UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Reduction and immobilization of selenium in wetland sediment Hodaly, Al Henry 2002

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2002-0110.pdf [ 12.16MB ]
Metadata
JSON: 831-1.0090217.json
JSON-LD: 831-1.0090217-ld.json
RDF/XML (Pretty): 831-1.0090217-rdf.xml
RDF/JSON: 831-1.0090217-rdf.json
Turtle: 831-1.0090217-turtle.txt
N-Triples: 831-1.0090217-rdf-ntriples.txt
Original Record: 831-1.0090217-source.json
Full Text
831-1.0090217-fulltext.txt
Citation
831-1.0090217.ris

Full Text

REDUCTION AND IMMOBILIZATION OF SELENIUM IN WETLAND SEDIMENT by AL HENRY HODALY B.Sc. Simon Fraser University 1999 A THESIS SUBMITTED IN PARTIAL FULLFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of Bio-Resource Engineering We accept this thesis as conforming to the required standard U N I V E R S I T Y ^ BRITISH COLUMBIA APRIL 2002 ©AL HENRY HODALY, 2002 In presenting this thesis in partial fulfillment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of /3/&- ffesooZctr f/Oa/^ft'R The University of British Columbia Vancouver, Canada Date flfir.2(> /?L£rO^ ABSTRACT Elevated levels of selenium have been found in stream waters downstream of surface coalmines in the Elk River Basin (McDonald and Strosher, 1998). The fate of Se in slow-moving side-channels and wetlands in the Elk River Valley is important because these types of ecosystems are some of the most important feeding and breeding habitats for fish and wildlife. The focus of this study was to observe the potential for selenium uptake in the sediments of these wetlands. If successful, wetlands could represent an effective passive remediation technology for the area. Sediment from Goddard Wetland was used in this study. Monitoring by the Elkview Mine found soluble selenium concentrations as high as 106 pg/L in the Goddard Tailing Pond, which feeds directly into Goddard Wetland. Sediment samples were inoculated into Thauera selenatis (T.sel) and sulphate-reducing bacteria (SRB) growth media. After approximately 3 days a bright-red precipitate (Se°) was seen in the T.sel medium, and after 13 days the SRB medium turned jet-black (FeS), confirming the presence of these microorganisms in the sediment. In both media, selenate concentrations decreased significantly (619 ppb -> 15 ppb and 364 ppb ->22 ppb, respectively) confirming the ability of these organisms to reduce selenate. Adsorption experiments performed with sterilized sediment showed that selenite adsorbed very strongly (78-100% removal from solution) to sediment, while selenate did not adsorb at all. A semi-continuous microcosm experiment was run to determine whether these selenium immobilizing/removal mechanisms would occur in an open system representative of the Goddard Wetland. Results clearly demonstrated that Se concentrations in solution decreased markedly after spiking; within one week, average concentrations dropped from 401 pg/L to 37 pg/L There was no significant difference between organically amended and unamended microcosms, possibly because the sediment used in these trials already had a high organic matter content (TOC = 26%; Walkley-Black C = 16%). iii TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iii LIST OF TABLES vi LIST OF FIGURES vii ACKNOWLEDGEMENTS viii 1. INTRODUCTION 1 2. BACKGROUND 4 2.1. CHEMISTRY OF SELENIUM 4 2.1.1. SELENIUM SPECIATION 7 2.2. REDUCTION OF SELENIUM BY MICROORGANISMS 12 2.2.1. ASSIMILATORY REDUCTION OF SE OXYANIONS 14 2.2.2. DISSIMILATORY REDUCTION OF SELENATE 17 2.2.3. CHARACTERISTICS OF SELENATE-REDUCING BACTERIA 19 2.3. VOLATILIZATION OF SELENIUM BY MICROORGANISMS 21 2.4. SELENIUM ADSORPTION TO SEDIMENT 27 2.5. BIOSTIMULATION 32 2.6. COAL MINING AND SELENIUM 33 2.7. TREATMENT WETLANDS 34 3. SITE DESCRIPTION, SURVEY, AND CHARACTERIZATION 37 3.1. MATERIALS AND METHODS 43 3.1.1. SAMPLE COLLECTION AND PRESERVATION 43 3.1.2. PRELIMINARY ANALYSIS OF SEDIMENT SAMPLES 45 3.2. SITE SURVEY AND ANALYSIS CONCLUSIONS 49 4. BATCH MICROCOSM STUDY 51 4.1. INTRODUCTION 51 4.2. MATERIALS AND METHODS 52 4.2.1. MEDIA COMPOSITION 54 iv Sulphate Reducing Bacteria (SRB) Postgate Medium C 54 T. Selenatis Growth Medium 55 4.2.2. EXPERIMENTAL METHODS 57 4.2.3. ANALYTICAL METHODS 59 Qualitative Monitoring 59 SEM and EDX analysis 60 Sulphate Analysis 61 Selenium Analysis (HG-AF) 64 4.3. RESULTS 68 4.3.1. QUALITATIVE ANALYSIS 68 Treatment 1 - Sediment and SRB Medium 68 Treatment 2 - Sediment and T. sel Medium 69 Treatment 3 - Autoclaved Sediment and SRB Medium 69 Treatment 4 - Autoclaved Sediment and T. sel Medium 70 Treatment 5 - SRB Medium 70 Treatment 6 - T. sel Medium 70 Treatment 7 - SRB Inoculum and SRB Medium 71 Treatment 8 - SeRB Inoculum and T. sel Medium 71 4.3.2. SEM AND EDX ANALYSIS 74 4.3.3. SULPHATE ANALYSIS 77 4.3.4. HG-AF SELENIUM ANALYSIS 79 4.4. DISCUSSION 83 Sulphate Analysis 87 S1M3 Selenium Analysis 88 4.5. BATCH MICROCOSM CONCLUSIONS 91 5. ADSORPTION 94 5.1. INTRODUCTION 94 5.2. MATERIALS AND METHODS 94 5.2.1. EXPERIMENTAL METHODS 95 5.2.2. ANALYTICAL METHODS 100 SELENIUM ANALYSIS (HG-AF) 100 5.3. RESULTS 100 V 5.3.1. HYDRIDE GENERATION-ATOMIC FLUORESCENCE SELENIUM ANALYSIS 100 5.4. DISCUSSION 107 5.5. ADSORPTION CONCLUSIONS 109 6. CONTINUOUS MICROCOSM STUDY 111 6.1 INTRODUCTION 111 6.2. MATERIALS AND METHODS .112 6.2.1. EXPERIMENTAL METHODS 114 6.2.2. ANALYTICAL METHODS 122 TOC 122 PH TESTING ....122 SE ANALYSIS (HG-AF) 123 6.3. RESULTS 123 TOC 123 PH MONITORING 124 HG-AF SELENIUM ANALYSIS 126 6.4. DISCUSSION.. 129 6.5. CONTINUOUS MICROCOSM CONCLUSIONS 134 7. CONCLUSIONS AND RECOMMENDATIONS 135 7.1. CONCLUSIONS 135 7.2. RECOMMENDATIONS 137 APPENDIX A - LECO CN 2000 APPLICATION NOTES 139 APPENDIX B - STATISTICAL ANALYSIS OF BATCH MICROCOSM 142 APPENDIX C - PRELIMINARY BATCH MICROCOSM RESULTS 146 APPENDIX D - PRELIMINARY BATCH MICROCOSM SULPHATE ANALYSIS 152 REFERENCES 153 VI LIST OF TABLES Table 1 - Common compounds of selenium 6 Table 2 - Physicochemical properties of elemental selenium 7 Table 3 - Selenium reduction half-reactions and potentials ..10 Table 4 - Microorganisms that volatilize selenium 25 Table 5 - Major species of Se at various redox potentials 28 Table 6 - Sorption parameters of selenite in sterilized sandy loam soil 31 Table 7 - Soluble S0 42 " and Se concentrations in settling pond decant 42 Table 8 - B.C. MOE preliminary analysis for Se and S0 42 ' 46 Table 9 - Results of carbon and nitrogen analyses 48 Table 10 - Batch microcosm treatment descriptions 53 Table 11 - Energy-dispersive x-ray (EDX) analysis 76 Table 12 - Concentrations of Se corresponding Na 2Se0 3 and Na 2Se0 4 98 Table 13 - Selenite adsorption trial results 101 Table 14 - Selenate adsorption trial results 102 Table 15 - Continuous microcosm factorial design 113 Table 16 - Treatment descriptions for all continuous microcosms 116 Table 17 - TOC and N analysis of vegetation amendment 124 Table 18 - pH monitoring of continuous microcosms 125 Table 19 - Individual continuous microcosm treatment Se concentrations 127 Table 20 - Average continuous microcosm treatment Se concentrations 127 vii LIST OF FIGURES Figure 1 - Potential-pH equilibrium diagram for the system Se-water 11 Figure 2 - Summary Diagram of Biological Cycling Occurring in Wetland 13 Figure 3a - The reductive portion of the biogeochemical cycle of Se 15 Figure 3b - Assimilatory and dissimilatory pathways 16 Figure 4 - Map of Elkview property and sample site 39 Figure 5 - Layout of Goddard settling pond and sample site #1 40 Figure 6 - Photographs of Goddard Settling Pond decant and Wetland 41 Figure 7 - Use of constructed dredge to sample sediment 44 Figure 8 - Standard calibration curves used for sulphate analysis 63 Figure 9 - Schematic configuration of HG-AF System 67 Figure 10 - Digital photograph monitoring of batch microcosm trial S1M3 72 Figure 11 - Scrapings of red precipitate 75 Figure 12 - Summary of sulphate analysis for batch microcosm S1M3 78 Figure 13 - Se measured in S1M3 treatments with SRB medium 80 Figure 14 - Selenium measured in S1M3 treatments with T.sel medium 82 Figure 15 - Adsorption experiment schematic for sample sediment S1A1 99 Figure 16 - Relationship of theoretical to control selenite concentrations 103 Figure 17 - Relationship of theoretical to control selenate concentrations 104 Figure 18 - Selenite concentrations of controls and treatments 105 Figure 19 - Selenate concentrations of controls and treatments 106 Figure 20 - Diagrams of continuous microcosm and carbon column 115 Figure 21 - Average continuous microcosm treatment concentrations 128 Figure C1 - Monitoring of batch microcosm trial S1M1 150 Figure D1 - Sulphate analysis of batch microcosm S1M1 152 ACKNOWLEDGEMENTS This project was primarily funded by an NSERC operating grant to my supervisor, Dr. Susan Baldwin; without her guidance this research would not have been possible. I would also like to thank Les McDonald and Mark Strosher (of the BC Ministry of the Environment) for all of their time, effort, and resources with regards to sample collection and analysis. Also, the cooperation of the Elkview Coal Mine was paramount to accessing the sample site, as well as obtaining water quality monitoring data for Goddard Pond. Rawandeep K. Gill, a Chemical and Biological Engineering undergraduate student, helped to prepare adsorption and continuous microcosm samples and perform selenium analysis on the HG-AF. Dhanesh Kannangara (Chemical and Biological Engineering), Carol Dyck and Keren Ferguson (Soil Science) are acknowledged for their assistance with analytical analysis. Jurgen Pehlke (Bio-Resource Engineering), and Peter Roberts and Graham Liebelt of the Chemical and Biological Engineering machine shop are acknowledged for their help with construction of the experimental apparatus. I would also like to thank the members of my thesis defense committee, Dr. Victor Lo and Dr. Ken Hall, for their consideration and availability. Lastly, I would like to thank my family for their continued support and motivation, and my wife, Leanne for encouraging me and always being there. 1 1. INTRODUCTION Selenium enters aquatic habitats from a number of anthropogenic and natural sources, including coal mining, coal-based power generation, irrigation waters, and various other industries (e.g., glass and electric component manufacturing, and paint coloring) (Bronzetti et al., 1993; Bennett, 1984). The burning and processing of coal is the principal source of global selenium environmental contamination (Dobbs et al., 1996; Bronzetti et al., 1993). Aquatic systems receiving inputs from fly ash piles or ponds associated with coal burning and production typically have elevated Se concentrations; most documented cases of Se contamination have been due to drainage from fly ash ponds or from irrigation water of Se-rich soils (USEPA, 1987; Lemly, 1987). In Canada, elevated concentrations of selenium in sediment, water, and biota may be connected to mining activities; for example one such area is the Elk River Basin in southeast British Columbia. Surface coal mining in the Elk River Basin is a significant source of waterborne selenium entering streams and rivers via small tributaries draining mine disturbed lands. This has resulted in elevated aqueous selenium concentrations of up to 25 times the current B.C. water quality criteria of 1 ug/L in streams and rivers below these contributing tributaries (McDonald and Strosher, 1998). 2 The fate of Se in slow moving side-channels and wetlands in the Elk River Valley is yet to be studied. Of particular interest are wetlands receiving Se-contaminated waters, such as Goddard Wetland. Selenium research in these lentic waterways is important because these types of ecosystems are some of the most important feeding and breeding habitats for fish and wildlife. Research in these ecosystems is also important because wetland characteristics and biogeochemistry may be used to optimize present day selenium remediation technology. The solubility of minerals containing Se, the complexing ability of solid and soluble ligands, methylation, volatilization, and microbiologically mediated oxidation-reduction reactions are all potential processes controlling Se concentration, mobility, and toxicity in both the aquatic and sedimentary wetland environment (Masscheleyn and Patrick, 1993). Therefore, the objective of my research was to study selenium cycling in Goddard Wetland sediment receiving Se-laden effluent from the Goddard settling pond. This study focused in particular on the potential for selenium adsorption and microbiological transformation in Goddard Wetland sediment. This objective was accomplished by performing several studies, each with its own sub-objective: 1) An initial site survey and characterization of the Goddard Wetland was performed. 3 2) Batch microcosm experiments were then performed to test for presence of selenium-reducing bacteria in the sediment. 3) Adsorption experiments were run to determine the potential of sediment to adsorb the selenium oxyanions, selenite and selenate. 4) A continuous microcosm experiment was performed to simulate real wetland conditions and observe the rate of selenate uptake by sediment. 5) Organic amendments were added to continuous microcosms to determine if organic content effects selenium uptake. After this introduction, this thesis progresses with Chapter 2, a general background describing the biogeochemistry of selenium speciation and cycling. Chapter 3 provides a site description, survey and characterization of the Goddard Wetland. Chapters 4, 5, and 6 outline the experiments that were performed using batch microcosms, adsorption trials, and continuous microcosms, respectively. Chapter 7 provides conclusions and recommendations for future studies. 4 2. BACKGROUND This chapter provides a review of: 1) selenium chemistry, speciation, and reduction processes, 2) the adsorption of selenium by sediment and soil, 3) organic amendment addition to enhance Se uptake/reduction by sediment, 4) other coal mining areas that have experienced problems with selenium, and 5) the use of treatment wetlands for remediation of mine sites, 2.1. CHEMISTRY OF SELENIUM Selenium (Se) is classified as a metalloid because it has chemical and physical properties that are intermediate between both metals and nonmetals. Se belongs in Group VIA of the periodic table, and is very similar to sulfur in its chemistry. Se has four oxidation states: +VI, +IV, -II, and 0. Commonly found species include the dissolved inorganic oxyanions selenate (Se042~) and selenite ( S e 0 32 ) , dissolved organic selenide (-II) bonded to carbon, and particulate selenium, including elemental selenium (Se°). In oxygen compounds it exists as the 4+ and 6+ oxidation states, and has a valence of 2- in combination with hydrogen or metals. Common Se compounds found in the environment are listed in Table 1. There are six stable isotopes of Se with varying degrees of abundance ( 7 4Se, 7 6 Se, 7 7 Se, 7 8 Se, 8 0 Se, 8 2 Se), and pure Se is allotropic; it exists as gray hexagonal, red monoclinic, and vitreous amorphous forms. Selenium is an antioxidant, making it a useful component of inks, mineral and vegetable oils, and lubricants (Haygarth, 1994). Other physicochemical properties of Se are summarized in Table 2. Table 1 - Common compounds of selenium (Haygarth, 1994). Name and Formula Where Found Selenides (-II), Se2" (i.e. FeSe2) Reducing environments, e.g. soils; forms metal complexes, highly immobile. Dimethylselenide (DMSe), (CH3)2Se Gas formed by volatilization from soil bacteria and fungi. Dimethyldiselenide (DMDSe), (CH3)2Se2 Gas formed by volatilization from plants. Dimethylselenone/ methyl methyl selenite, (CH 3) 2Se0 2 Volatile metabolite, possibly formed as a final intermediate prior to reduction to DMSe. Hydrogen selenide, H2Se Gas, unstable in moist air; decomposes to Se in water. Elemental selenium (0), Se° Stable in reducing environments; (a) red crystalline alpha and beta monoclinic; (b) red glossy or black amorphous forms, all insoluble in water and oxidation/reduction; very slow. Selenite (+IV), Se0 32" Soluble form, common in mildly oxidizing conditions, e.g. soils or air particles. Trimethylselenonium (TMSe), (CH3)3Se+ Important urinary metabolite of dietary Se and is made rapidly unavailable to plants by fixation and volatilization. Selenous acid, H2Se03" Selenous acid is protonated in acid/neutral conditions. Se(IV) is easily reduced to Se° by ascorbic acid (Vitamin C) or sulfur dioxide in acidic environments by microorganisms. Readily available by Fe oxides, amorphous Fe hydroxides, and aluminum sesquioxides in soils. Selenium dioxide, Se0 2 Hse03" Gas formed as a product of fossil fuel combustion (sublimation temperature = 300°C); Dissolves in water to form selenous acid. Common soils. Selenate (+VI), Se0 42" Se(VI) is stable in well-oxidized environments, and very mobile in soils, hence easily available to plants. Slowly converted to more reduced forms; not as strongly absorbed as Se(IV). Selenic acid, H 2Se0 4, HSe0 4 ' Common in soils. 7 Table 2 - Physicochemical Properties of Elemental Selenium (Haygarth, 1994; Newland, 1982). Atomic number 34 Atomic mass 78.96 Density, g/cmJ 4.79 (hexagonal modification) Melting point, °C 217 Boiling point, °C 685.4 Atomic radius, pm 0.117 Hardness, relative units 2 (hexagonal modification) Electronegativity, relative units (Li=1) 2.4 Latent heat of fusion, J/g (Cal/g) 6.91 (16.5) Heat of vaporization, J/g (Cal/g) 272.98 (65.2) Thermal conductivity, W (m °C) 0.293 - 0.766 2 .1 .1 SELENIUM SPECIATION Eh-pH diagrams show how protons and electrons shift equilibria under various conditions, and can indicate which species of Se predominate under any given condition of pH and redox potential (or Eh). These diagrams are constructed using the Peters-Nernst equation: = Eh° + (2.3 RT/nF) log [(oxidant)/(reductant)] Where Eh° = pe° (2.3 RT/F) R = 8.314 Jmor 1rC 1 T = 25°C (298.15 K) standard conditions with P = 1 atm F = 96 490 C mol"1 electric charge of 1 mol of electrons n = number of electrons transferred pe = log K ( K = [products]/[reactants]) = -AG°/2.3RT (AG° = Gibbs free energy) AG° = - R T I n K substituted into the above equation we find. ps° = In K / 2.3 Reduction potential equations in Table 3 are derived from the Peters-Nernst equation above. The reduction potentials for these half-reactions at 9 different pH's can be used to construct a potential-pH equilibrium diagram (Figure 1) for the system selenium-water (at 25°C). This diagram is only valid in the absence of substances with which selenium can form insoluble compounds or soluble complexes. Figure 1 indicates that selenium is a fairly noble substance because a large portion of its stability domain covers that of water. It is therefore stable in water and aqueous solutions of all pH's free from oxidizing and reducing agents (Van Muylder and Pourbaix, 1994). 10 Table 3 - Selenium reduction half-reactions and potentials (Van Muylder and Pourbaix, 1994). Equation Reduction Potential (E h) 1. Selenate (+6) -> selenite (+4) HSe0 4~ + 3 H+ + 2e" -> H 2 S e 0 3 + H z O E h = 1.090- 0.0886 pH + 0.0295 log (HSe0 4 ") / (H 2 Se0 3 ) S e 0 42 " + 4 H + + 2e" -> H 2 S e 0 3 + H 2 0 E h = 1.151 - 0.1182 pH + 0.0295 log (Se0 42 " ) / (H 2 Se0 3 ) Se0 42 " + 3 H + + 2e" Hse0 3 " + H z O E h 1.075- 0.0886 pH + 0.0295 log (Se0 42")/( HSe0 3") S e 0 42 " + 2 H + + 2e~ -> S e 0 32 " + H 2 0 E h = 0 . 8 8 0 - 0.0591 pH + 0.0295 log (Se0 42 " ) / (Se0 32 ) 2. Selenite (+4) -» selenide (-2) H 2 S e 0 3 + 6 H+ + 6e" -> H 2Se + 3 H 2 0 E h = 0.360 - 0.0591 pH + 0.0098 log (H 2 Se0 3 ) / (H 2 Se) HSe0 3 " + 7 H+ + 6e" -> H 2Se + 3 H 2 0 E h = 0 . 3 8 6 - 0.0690 pH + 0.0098 log (HSe0 3")/(H 2Se) HSe0 3 " + 6 H+ + 6e" -> HSe" + 3 H 2 0 E h = 0.349 - 0.0591 pH + 0.0098 log (HSe03")/(HSe") Se0 32 " + 7 H + + 6e" -> HSe" + 3 H 2 0 Eh = 0 . 4 1 4 - 0.0690 pH + 0.0098 log (Se0 32")/(HSe") Se0 32 " + 6 H + + 6e" -> Se 2" + 3H z O E h = 0 . 4 1 4 - 0.0690 pH + 0.0098 log (Se0 32")/(Se 2") 3. Selenite (+4) ^ Se (0) H 2 S e 0 3 + 4 H+ + 4e" -> Se ( S ) + 3 H 2 0 E h = 0.741 - 0.0591 pH + 0.0148 log (H 2 Se0 3 ) HSe0 3 " + 5 H+ + 4e" -> Se ( S ) + 3 H 2 0 Eh = 0 . 7 7 8 - 0.0739 pH + 0.0148 log (HSe0 3") S e 0 32 " + 6 H + + 4e" -> Se ( S ) + 3 H 2 0 E h = 0 . 8 7 5 - 0.0886 pH + 0.0148 log (Se0 32 ") 4. Selenide (-2)-> Se (0) H 2Se -> Se ( S ) + 2 H+ +2e" E h = -0.399 - 0.0591 pH - 0.0295 log (H 2Se) HSe" -> Se ( S) + H+ +2e" E h = -0.510 - 0 . 0 2 9 5 pH -0.0295 log (HSe) Se 2" -> Se ( S ) + 2e" Eh = - 0.924 - 0.0295 log (Se2") Figure 1 - Potential-pH equilibrium diagram for the system selenium-water, at 25°C (Van Muylder and Pourbaix, 1994). Although redox potential was not measured in Goddard Wetland, the area within the oval indicates the pH and suspected redox potential of a lentic system with standing water year-round. HSe04" 1 1 Se0 4 2 _ ^^HaSeOa 1 "* — -» 1—• . HSeO, ** - ^ / \ — . | Se ^ S e 0 3 2 1 I H2Se 1 ' 1 1 1 1 1 \ HSe - / 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 PH When considering the redox stability of water (shaded in figure 1), values above the shaded area show the conditions for which water is oxidized to 0 2 . Accordingly, values below the shaded area indicate conditions for which water is reduced to H2. 12 The selenates are stable in well-oxidized environments, and very mobile in soils, and are therefore quite easily available to plants. Selenic acid (HSeCV) is predominant at very low pH's (Figure 1). At the opposite end of this figure, in reducing environments, the selenides are predominantly found. These may be found in soils and sediments, they can form metal complexes such as iron selenide, and are highly immobile. Elemental selenium, is stable in reducing environments, insoluble in water, and is often found in the form of a bright red precipitate. Lastly, the stability domain of the selenites largely covers that of water (Figure 1). Selenite is the most soluble form of Se, and is common in mildly oxidizing conditions (i.e. soils). 2.2 REDUCTION OF SELENIUM BY MICROORGANISMS Figure 2 illustrates the many mechanisms that govern selenium cycling in aquatic systems. For this study, we have focused on processes occurring in sediment that are capable of immobilizing and/or removing selenium from these systems. These processes include adsorption to sediment, and reduction to immobile or potentially volatile forms of Se. Microorganisms can reduce selenium (as selenate and/or selenite) to elemental Se° and gaseous Se2". These reduction and volatilization reactions (refer to Figure 2) are discussed in the following sections (2.2 and 2.3). Figure 2 - Summary Diagram of Biological Cycling Occurring in Wetland (USEPA, 1998). 13 Reduction Methylation Adsorption Microorganisms are known to have biochemical interactions with selenium that can affect its chemical speciation and complexation (Doran, 1982) and therefore may be significant in affecting its mobility in nature. Because many of the environments that receive seleniferous waste can be characterized as anoxic (i.e. subsurface saturated soils or organic-rich marsh sediment), biochemically-14 mediated transformations of selenium carried out by anaerobic bacteria have been a logical area of investigation. This section summarizes the assimilatory reduction of Se oxyanions, the dissimilatory reduction of selenate, and characteristics of selenate-reducing bacteria. 2.2.1. ASSIMILATORY REDUCTION OF SE OXYANIONS Selenium and sulfur are neighbors in the periodic table as Group VIA elements and hence share a great similarity of chemical and biochemical properties. The toxicity of selenium oxyanions appears related to their mimicry of sulfur oxyanions. This may be relevant to understanding the chemical speciation and biogeochemical cycling of selenium in nature (Cutter and Bruland, 1984). In many microbes, selenate is transported into cells by sulfate permeases, while in some species, the transport pathway for selenite is distinct from the one for selenate/sulfate or sulfite (Brown and Shrift, 1980; Bryant and Laishley, 1988). Selenite and selenate, upon entering cells, undergo assimilatory reduction to the level of selenide, at which point they may be incorporated into protein or released as alkylselenides (Wrench, 1978; Karle and Shrift, 1986). Therefore, reduced organosulfur compounds, such as methionine, can also alleviate selenium toxicity. One common observation of uptake and assimilatory 15 reduction studies (Doran, 1982) has been the precipitation of red elemental selenium (Se°) from bacterial cells exposed to selenite. Although reasons for this phenomenon are unclear, it is known that selenite is readily reduced to the elemental state by chemical reductants such as sulfide or hydroxylamine or biochemically by systems such as glutathione reductase (Rashid and Krouse, 1985). This precipitation reaction is also associated with bacterial dissimilatory reduction (Figure 3a,b) and may therefore have great environmental significance. Figure 3a - The reductive portion of the biogeochemical cycle of selenium (Oremland, 1994). BACTERIAL DISSIMILATORY REDUCTION Selenate (Se+6) Selenite (Se+4) BIOLOGICAL REDUCTION Mt Mm CHEMICA ^ REDUCTH ADSORPTION L ION Elemental Selenium (Se°) ASSIMILATION Selenides Organoselenium BACTERIAL ASSIMILATORY REDUCTION METHYLATION Volatile Selenides (e.g., DMSe) DEMETHYLATION C 0 2 + C H 4 16 Figure 3b - Assimilatory and dissimilatory pathways and possible precursors for the formation and destruction of volatile alkylated sulfur and selenium compounds (Oremland, 1994). Sulphate Selenate A S S I M I L A T O R Y R E D U C T I O N Methionine C H J S C H J C H J C H C O O -Selenomethionine CH 3 SeCH 2 CH 2 CHCOO DECOMPOSITION Methanethiol CH 3 SH 7 Methylation/ Condensation H O G H ? C H 2 C H C O O -Methanehydrogen -selenlde CH 3SeH Methylation/ Condensation •PMSP S C H , C H , C O O -H , C • DMSeP. SeCHjCHgCOG-JL DECOMPOSITION Dimethylsutfide H 3 C-S^CHj Oxidation Aaodeis C H 2 - CHCOO-Dimethylaelen ids H 3 G - S e - C H 3 Dimethyldieulfide H3C--S—S—CH3 Oxidation Dimethyldiaelenide H 3 C - S ® - S e - C H 3 BACTERIAL CATABOLISM V An, A aerobic C 0 2 + S0 4 ~ gr, SeO^ " 17 2.2.2 DISSIMILATORY REDUCTION OF SELENATE The possibility that microorganisms involved in the sulfur cycle could also achieve reductive transformation of traces of selenium was investigated by Zehr and Oremland (1987). The rationale was that since the sulfate/selenate ratio in Se-impacted environments was typically about 1000:1 (Oremland et al., 1989), perhaps some selenate could cycle through the reductive systems operative for sulfate. Sulfate-reducing bacteria are inhibited by oxyanions of Group VI, such as selenate (Postgate, 1949; 1952; Peck, 1959). These anions form unstable analogues of adenosine-51-phosphosulfate, the product of the first reaction of the sulfate reduction pathway involving ATP-sulfurylase. This thermodynamically unfavorable reduction requires the investment of a molecule of ATP: S0 42 " + ATP APS + pyrophosphate The possibility was indicated for reduction of trace concentrations of selenate to selenide (e.g. picomolar-nanomolar) via the pathway for sulfate reduction (Zehr and Oremland, 1987). This reduction was shown to occur with washed cell suspensions of Desulfovibrio desulfuricans, which reduced trace quantities (= 70 picomolar) of selenate to selenide. Even with the manipulation of bacterial cells and sediment, only an insignificant fraction (<0.01) of selenate was reduced to selenide in anaerobic sediments. However, during the course of this study, a novel dissimilatory reduction of selenate to Se° was observed in 18 sediment porewater profiles of selenate and sulfate taken from an agricultural wastewater evaporation pond. Whereas sulfate was consumed at depth by bacterial reduction, selenium oxyanions were consumed by a process occurring near the sediment surface and in the presence of extremely high concentrations (= 320mM) of sulfate. In the laboratory, anaerobic sediment slurries consumed millimolar quantities of selenate, and this was enhanced by provision of electron donors such as acetate, hydrogen, or lactate. This reduction proceeded with selenite as a transient intermediate, followed by the accumulation of a red precipitate (discussed previously). No volatile gases (i.e. DMSe) or selenide precipitates (i.e. FeSe) were detected. The product of this reduction was elemental selenium (Se°). The significance of dissimilatory selenate reduction (DSeR) in nature was assessed by the development of an in situ radioassay that measures the precipitation of Se° into sediment from injected selenate (Oremland et al., 1989). The assay was applied to the sediments from an agricultural evaporation pond in the San Joaquin Valley and demonstrated that along with denitrification, DSeR was confined to the surficial (top few centimetres of) sediments while sulfate reduction occurred at greater depth. The potential for DSeR and denitrification extended down the length of the sediment core sample. The areal rate of DSeR was calculated to be 0.3 mmol/m2/day, which was about a factor of 3 lower than nitrification and about a factor of 30 lower than that for sulfate reduction. Nonetheless, the turnover time for selenate in the water column was 82 days, 19 which was markedly more rapid than that for nitrate (2009 days) or sulfate (49, 197 days). These results indicate that immobilization of selenate to Se° is a practical approach for Se remediation. 2.2.3 CHARACTERISTICS OF SELENATE-REDUCING BACTERIA The microbiology and biochemistry of selenate-respiring (or reducing) bacteria, has been studied in only a few laboratories. The precipitation of Se° from selenate was reported in cultures recovered from the Kesterson National Wildlife Refuge (Maiers et al., 1988), although the mechanisms for this were not pursued. Isolation of an anaerobic coccus (strain SES-1) that grew in mineral medium with acetate as the electron donor and selenate as the electron acceptor was first reported by Oremland et al. (1989). The extent of acetate oxidation was dependent upon the availability of selenate, and a stoichiometric balance was achieved between selenate reduced, Se° recovered, and acetate oxidized (Oremland, 1994). This dissimilatory reduction was carried out according to the reaction: 4 CH3COO" + 3 Se0 42" -» 3 Se° + 8 C0 2 + 4 H 2 0 + 4 H+ 20 Macy etal. (1989) reported the isolation of a pseudomonad that grew by oxidation of acetate with reduction of selenate to selenite. Elemental selenium was precipitated by the reduction of selenite carried out by a second organism, a gram-positive rod. The thermodynamics of selenate reduction to selenite is highly favorable (Oremland et al., 1989; Macy et al., 1989) and contrasts with the endergonic reaction associated with the reduction of sulfate to sulfite (Steinberg et al., 1992). Rech and Macy (1992) reported distinct terminal reductases for nitrate and selenate respiration for their isolate, which they no longer termed a pseudomonad but named Thauera selenatis. Steinberg et al. (1992) screened five common laboratory cultures of pseudomonads (e.g. Pseudomonas stutzeri) and Halobacterium denitrificans for the ability to grow anaerobically using selenate as an electron acceptor. None of the cultures grew or achieved chemical reduction of selenate. However, Lortie et al (1992) reported isolation of a wild-type strain of P. stutzeri that was capable of vigorous reduction of selenate to Se° during aerobic growth. Reduction was inhibited by chromate and tungstate but not by nitrate, nitrite, or sulfate. Similar results were reported by Barnes et al. (1992) with an isolate identified as P. stutzeri. All of these observations suggest that selenate reduction in pseudomonads is associated only with novel isolates that have been taken from Se-contaminated environments and grown with selenate in the medium. Tomei et al. (1992) reported that an adapted selenium-resistant strain of Wolinella succinogenes could reduce selenate (and selenite) to Se° during the stationary 21 phase after anaerobic growth, but it could not grow using selenium oxyanions for respiration. All of the above reports indicate that selenate reduction is associated with detoxification mechanisms rather than dissimilatory reduction (Oremland, 1994). 2.3 VOLATILIZATION OF SELENIUM BY MICROORGANISMS Volatilization through methylation is thought to be a protective mechanism used by microorganisms to avoid Se toxicity in seleniferous environments. This process permanently removes Se from soil and water under aerobic conditions (Frankenberger and Karlson, 1994). In water bodies, bacteria are thought to play a dominant role in the biomethylation of toxic Se species such as Se0 32" and Se0 42" into the less toxic, volatile form DMSe (Thompson-Eagle and Frankenberger, 1990). Following methylation into this volatile species, it is released into the atmosphere, diluted, and dispersed by air currents away from the contaminated source. Table 4 summarizes the various microorganisms that volatilize selenium. 22 There are a number of factors that enhance the volatilization of selenium in both soil and water: Microorganisms - bacteria and fungi are important in this process. For example, adding a fungal innoculant of 2.8 x 107 cells of Candida humicola per gram of soil caused Se evolution to double (Zieve and Peterson, 1981). Nutrients - Soil alkylselenide production is often carbon limited. In general, the rate of Se evolution from soils, sediments, and water increases with the addition of organic amendments (Frankenberger and Karlson, 1994). It is possible to achieve a 10-fold increase in volatile Se evolution with the addition of organic amendments to soil. Effective treatments for accelerating volatilization from seleniferous soils are gluten, casein, pectin, and citrus peel (Calderone et al., 1990; Frankenberger and Karlson, 1989; Karlson and Frankenberger, 1990). Selenium concentration - Although the Se volatilization capacity of a soil is dependent on Se concentration in the soil (Karlson and Frankenberger, 1988), it is the level of available or water-soluble Se that controls this process. Zieve and Peterson (1981) were able to correlate a decrease in water-soluble Se with decreasing methylation rates as volatilization proceeded. Also, Reamer and Zoller (1980) reported that the relative abundance of volatile Se species evolved from Se0 32" upon the addition of sewage sludge was dependent on the Se concentration (Frankenberger and Karlson, 1994). 23 Selenium species - Doran (1982) determined that the pathway of Se biomethylation requires reduction to the Se2" species and subsequent methylation to form DMSe. It should therefore be more favorable energetically to methylate Se0 32" rather than Se0 42". Temperature - Selenium volatilization is temperature-dependent. The maximum release of DMSe from lake sediments occurred at 20°C (Chau et al., 1976), in California evaporation pond water at 35°C (Thompson-Eagle and Frankenberger, 1990), from a loamy soil at 20°C (Zieve and Peterson, 1981), and from a California sandy-textured seleniferous soil at 35°C (Frankenberger and Karlson, 1989). However, it is possible the optimum temperature for Se volatilization may not have been reached in these studies because maximum DMSe emission occurred at the maximum temperature tested. Moisture - Air-drying the soil severely inhibits Se volatilization (Zieve and Peterson, 1981), while water-saturating it causes anaerobiosis, which also decreases production of volatile Se (Frankenberger and Karlson, 1989). Field studies have shown that Se emission rates are much lower at dryer sites than in corresponding damp or wet conditions (Frankenberger, 1989; Frankenberger and Karlson, 1988; Frankenberger and Karlson, 1989). 24 pH - The biologically mediated biomethylation process as well as the solubility and availability of Se are affected by pH. The optimum pH for Se biomethylation and subsequent volatilization from a seleniferous Kesterson sediment (pH 7.7) was 8.0 (Karlson and Frankenberger, 1989). The addition of lime to a sandy soil increased the pH from 6 to 7 and increased Se volatilization 1.2-fold (Hamdy and Gissel-Nielson, 1976). Aeration - Several studies have shown that greater quantities of volatile Se are evolved under aerobic conditions than under anaerobic conditions (Abu-Eirreish et al., 1968; Thompson-Eagle and Frankenberger, 1990; Francis et al., 1974; Reamer and Zoller, 1980). However, substantial evolution of volatile Se from seleniferous soils has also occurred under anaerobiosis (Doran and Alexander, 1977; Karlson and Frankenberger, 1993). Factors which are inhibitory to the volatilization of selenium include the presence of heavy metals such as molybdenum, mercury, chromium, and lead (Karlson and Frankenberger, 1988), and the presence of high levels of nitrate and nitrite (Oremland etal., 1989; Karlson and Frankenberger, 1988; Thompson-Eagle and Frankenberger, 1991). 25 Table 4 - Microorganisms that volatilize selenium. (Frankenberger and Karlson, 1994). Reference Organism Source / Environment Se Substrate Se cone, (ppm) Se Product Bacteria Chauetal., 1976 Aeromonus sp. Lake sediment / Aerobic Se0 3 5 (CH3)2Se Flavobacterium sp. Lake sediment / Aerobic Se0 3 5 (CH3)2Se2 Pseudomonas sp. Lake sediment / Aerobic Se0 3 5 Unknown volatile Se Doran and Alexander, 1975 Corynebacterium sp. Seleniferous soil / Aerobic Se0 3, Se0 4, Se° - (CH3)2Se Chasteen et al., 1990 Pseudomonas fluorescens Evaporation pond sediment / Anaerobic SeO„ 0.8 (CH3)2Se, (CH3)2Se2 Thompson-Eagle and Frankenberger, Unpub. Unidentified sp. Evaporation pond sediment/Aerobic Mainly Se0 4 1.2 (CH3)2Se Fungi Barkes and Fleming, 1974 Cephalosporium sp. Fusarium sp. Garden soil / Aerobic Garden soil / Aerobic Se0 3 Se0 4 457 418 (CH3)2Se (CH3)2Se Penicillium cithnum Garden soil / Aerobic Se0 4 418 (CH3)2Se Scopulariopsis sp. Garden soil / Aerobic Se0 4 418 (CH3)2Se Challenger and North, 1934 Scopulahopsis brevicaulis Unspecified / Aerobic Se0 3, Se0 4 15 (CH3)2Se Challenger and Charlton, 1947 Schizopyllum commune Wood / Aerobic Se0 4 366 (CH3)2Se 26 Table 4 continued... Reference Organism Source / Environment Se Substrate Se cone, (ppm) Se Product Challenger, 1951 Aspergillus niger Unspecified / Aerobic SeO„ - (CH3)2Se Cox and Alexander, 1974 Candida humicola Sewage / Aerobic Se0 3, Se0 4 46 (CH3)2Se Candida humicola Sewage / Aerobic Se0 3, SeO, 418 (CH3)2Se Karlson and Frankenberger, 1988 Acremonium falciforme Evaporation pond sediment / Aerobic 7 5 Se0 3 100 (CH3)2Se Penicillium citrinum Ulocladium tuberculatum Evaporation pond sediment / Aerobic Evaporation pond sediment / Aerobic 7 5 Se0 3 7 5 Se0 3 100 100 (CH3)2Se (CH3)2Se Chasteen et al., 1990 Acremonium falciforme Evaporation pond sediment / Aerobic Se0 4 0.79 (CH3)2Se Penicillium citrinum Evaporation pond sediment / Aerobic SeO„ 0.79 (CH3)2Se2 Fleming and Alexander, 1972 Penicillium sp. Sewage / Aerobic Se0 3 457 (CH3)2Se Thompson-Eagle et al., 1989 Alternaria altemata Evaporation pond water / Aerobic Se0 3 1 (CH3)2Se Altemaria altemata Evaporation pond water / Aerobic Se0 4 100 (CH3)2Se 27 2.4. SELENIUM ADSORPTION TO SEDIMENT In addition to reduction and volatilization of selenium, immobilization can occur via adsorption (binding and complexation) onto clay and the organic phase of particulates (Lemly, 1987). Stable selenite ions are able to migrate until they are adsorbed on mineral or organic particles. In consequence, the selenite level is increased in several coals as well as in clay sediments; this may explain the origin of selenium in coals. Frost and Griffin (1977) have found that selenites are the preferable species of Se being adsorbed by clay minerals, particularly by montmorillonite and Fe oxides. Weerasooriya et al. (1989) have also found that the adsorption of selenite by goethite is highly pH dependent (most readily adsorbed at a pH range of 2.0-3.0). Elrashidi et al. (1987) studied equilibrium reactions and constants for 83 Se minerals and solution species that relate to soils and observed that the redox (pe + pH) of soils controls Se speciation in solution (Table 5). They described Se species and Se-metal complexes that are likely to occur in soils under various conditions. In soils of high Ca and Mg concentrations, both CaSe0 4 and MgSe0 4 contribute to the total Se in solution, whereas in acidic and very acidic soils the most important compounds seem to be KHSe. NH4HSe, and MnSe. 28 Table 5 - Major species of Se at various redox potentials (Elrashidi et al., 1987). Redox Value (pe + pH) pH value Major species in soil solution High (17) 7 Se<V" High (17) <2 HSeOV , H 2Se0 3 Moderate (12) >7 Se03"" Moderate (12) <7.3 HSe03" Low (4) >3.8 HSe Low (4) <3.8 H2Seu Several studies (Lakin and Dawidson, 1967; Allaway, 1968; Paasikallio, 1981; and Combs, 1986) have extensively reviewed Se behavior in soil and commented on its complexity (Kabata-Pendias, 1991). They have generalized that: 1. In acidic gley (or sticky clay) soils, and soils with high organic matter content, selenides and Se-sulfides dominate - they are only slightly mobile and therefore not very available to plants. 29 2. In well-drained mineral soils with pH close to neutral, selenites exist exclusively. Their alkaline metal compounds are soluble, but Fe selenites are not; moreover, selenites are rapidly and nearly completely fixed by Fe hydroxides and oxides and thus are very slightly available to plants. 3. In alkaline and well-oxidized soil selenates are likely to occur. They are easily soluble and are unlikely to be fixed by Fe oxides and may be highly mobile and readily taken up by plants. Guo et al. (1999a) conducted batch sorption experiments to characterize selenite sorption in soil. The sorption of selenate was not measured due to its reported negligible adsorption irrespective of all soils (refer to Chapter 5 discussion). Guo et al. found that adsorption isotherms for selenite were all nonlinear with an obvious curvature at equilibrium concentrations above 1.5 pg/mL The adsorption behaviour of selenite can be expressed by the empirical Freundlich equation: S = K/ C" Where, S is the surface bound concentration of selenite, C is the concentration of selenite in solution, and Kf and n are constants. 30 Guo et al. determined adsorption parameters for three different soil treatments (Table 6). They found that the severe nonlinearity of selenite adsorption was manifested by the values of Freundlich exponent n, which all significantly deviated from unity (0.69-0.73). The adsorption capacity, as measured by Freundlich constant Kf, decreased when amended with manure and increased when amended with gluten. This was also true for the linearized distribution coefficient Kd, when a simple linear relationship between C and S was assumed and applied to adsorption isotherms. These adsorption coefficients (Table 6) indicate that selenite is adsorptive in soil despite its anion forms (SeO*32"). Sposito et al. (1988) described this type of selenite sorption in soil as ligand exchange reactions involving the formation of inner-sphere surface complexes between SeC»32" and reactive surface hydroxyl groups. Guo et al. (1999a) found that adsorption was not increased by amendment with manure; this is consistent with the mechanism proposed by Sposito et al (1988). Also, the increase in adsorption by gluten is likely due to the presence of amino groups on the protein, which could provide positive charge sites upon protonation and thus could sorb Se0 32" through outer-sphere surface complexation (Sposito, 1989). 31 Table 6 - Sorption parameters of selenite in sterilized sandy loam soil (Guo et al., 1999). Soil Treatment Batch-Derived Column-Derived Kd, mL/g n K,, mL/g r* K d , mL/g Unamended 0.726 2.120 0.987 1.631 0.335 Gluten-amended 0.715 3.042 0.984 2.343 1.315 Manure-amended 0.691 0.881 0.996 0.651 0.198 32 2.5. BIOSTIMULATION Research on the effect of organic amendments to selenium uptake by sediment is limited in scientific literature. As mentioned in the previous section, Guo et al. (1999a) found that addition of gluten decreased the mobility of selenium, where 72% of Se was retained in the soil in the form of elemental selenium or as further reduced (Se2") forms. Much of the literature has observed the effects of organic amendments on selenium volatilization from soils. Some of these studies are discussed in Chapter 2.3, and are summarized below. Stork et al. (1999) tested two amino acids (DL-homocyteine and L-methionine), a carbohydrate (pectin), and a protein (zein) on several soils. Both zein and L-methionine were found to strongly increase volatilization (up to 43% of applied Se), whereas DL-homocysteine had a much smaller stimulating effect. Pectin had a moderate effect, but enhanced Se volatilization rates were sustained much longer when compared to the zein amendment. Other studies (Calderone et al., 1990; Frankenberger and Karlson, 1989; Karlson and Frankenberger, 1990; Zhang and Moore, 1997a; Zhang and Frankenberger, 1999) have observed the effects of organic amendments on selenium volatilization from soils. These amendments were found to increase volatilization from seleniferous soils: gluten, casein, casamino acids, pectin, citrus peel, and decomposed wetland plants. 33 2.6. COAL MINING AND SELENIUM Selenium mobilization as the result of surface coal mining in Wyoming, Alberta, and British Columbia have only recently come to light in the 1990s (Dreher and Finkelman, 1992; Casey and Siwik, 2000; McDonald and Strosher, 1998). In these cases, the actual sources of selenium are still being explored, however they appear to be associated with selenium-bearing rock and soil, and various activities associated with surface mining. A survey of the McLeod, Pembina, and Smoky Rivers in Northern Alberta has indicated that selenium concentrations from surface waters downstream of the CRC, Gregg River and Smoky River coal mines exceed the current C C M E guideline of 1 pg/L, as well as the US-EPA chronic guideline of 5 pg/L (Casey and Siwik, 2000). Further investigation of sediment from Lac des Roches (near the Cadomin coal mine) has shown consistently higher selenium concentrations (approximately 3-50 times higher) than sediment sampled from the reference site, Fairfax Lake. Mean selenium concentrations in rainbow trout muscle and eggs ranged from 0.13 to 9.34 pg/g wet weight in muscle and from 0.02 to 28.90 pg/g wet weight in eggs. These concentrations showed a positive correlation with selenium concentration in water samples taken about the same time. Our research has focused on selenium mobilization from Elkview Coal mines in southeast British Columbia. Chapter 3 of this study provides an in-34 depth discussion of how selenium mobilization from open pit mining in the Elk River Basin has contributed to elevated concentrations of selenium in the adjacent Goddard Wetland. 2.7. TREATMENT WETLANDS As a result of surface coal mining, an increase in selenium mobilization will impact on: i) the environment - elevated levels of selenium may negatively impact both fish and birds. An evaluation of monitoring and assessment studies by Lemly (1987) recommends that waterborne selenium concentrations of 2 ug/L or greater be considered highly hazardous to the health and long-term survival of fish and wildlife. ii) mining companies - elevated concentration of selenium in the environment may lead to clean-up costs and the highly expensive introduction / maintenance of a chemical treatment facility, or the closure of productive mine sites. 35 Because of these impacts, it would be beneficial to find natural processes that might mitigate the problem of selenium in water, sediment and biota. A possible solution is the utilization of enhanced or constructed wetlands and/or ponds that make use of reduction, volatilization, and/or bioremediation processes described in this study. The design, construction and operation of a chemical treatment plant is a very expensive method. The open-pit copper-molybdenum Brenda Mines, 22-km Northwest of Peachland, British Columbia, closed in 1990. In 1998, a 912-cubic meter per hour treatment facility was designed and constructed to treat elevated concentrations of molybdenum being discharged from the old mine site. This facility cost the Brenda Mines, a division of Noranda Inc., an estimated $11 million to design and construct (Stroiazzo, 1999). We believe that a much better alternative to an expensive chemical treatment facility is the use of natural processes, such as the reduction, volatilization, and bioremediation phenomena occurring in wetlands. The use of constructed wetlands in treating selenium is becoming increasingly popular. It is a method of treatment that is aesthetically pleasing to the eye, and is accompanied by costs that are orders of magnitude lower than other treatment systems (Hansen etal., 1998; Kadlec and Knight, 1996). The information obtained from our research may be used to optimize the efficiency of treatment wetlands and/or ponds that make use of reduction, volatilization, and bioremediation processes (such as adsorption). Lastly, this research may impact the performance and potential of wetlands to be used as treatment systems for the future. 37 3. SITE DESCRIPTION, SURVEY, AND CHARACTERIZATION Large scale surface coal mining in the Elk Valley commenced in 1970 and has since expanded in production and area disturbed. The Elkview mine, located 15-km from Sparwood in southeastern British Columbia (Figure 4), has steadily increased production since 1970 to its current level of four million tonnes of metallurgical coal. The coal occurs in 11 seams that vary in thickness from 2 to 15-meters. The existing coal reserve would sustain operations for over 40 years at current mining rates. A conventional shovel/truck operation is used in the open pits. McDonald and Strosher (1998) have confirmed that significant quantities of selenium, mostly in the dissolved form, are being mobilized by surface coal mining into the Elk River system. The mobilization of soluble forms of selenium may be most detrimental to slower moving side channels and wetlands in the Elk Valley. The Goddard wetland receives drainage water directly from the Goddard settling pond on Elkview mine property. The wetland was rather expansive, therefore a site survey was first performed and several different areas sampled; these areas include three shallow ponds (Sites #1, 2, 3), one deeper beaver pond (Site #4), and the Goddard Settling Pond inlet (Site #5). Samples were 38 collected from three different locations in each pond so as to obtain a representative sample. Site #1 (discussed in further detail later) was a small, shallow pond closest to the settling pond decant. This site was chosen for the work described in this thesis based on its proximity to the Goddard Settling Pond, as well as the direct flow of settling pond drainage through the site (Figures 5 and 6a,b). Site #1 was approximately 10 by 3 meters in area, and surrounded by common cattail, Typha latifolia, and beaked sedge, Carex rostrata. Data and sample collection was taken on the morning of April 25, 2001, at which time the water temperature was 10.7°C, water pH was 9.5, and water depth above sampled sediment ranged from 11.5 - 15 cm. A preliminary survey and characterization of water flow through Goddard Wetland indicated that the settling pond decant was the main source, of water flow through Site #1. Water quality data for sulphate and total selenium was monitored periodically from April 1996 to April 2001 from the Goddard Pond decant; this data can be found in Table 7. This data indicates that the Goddard Settling Pond has a history of high selenium concentrations, in which values collected during the monitoring period exceeded the CCME guideline (1 ug/L) by 2- to 100-times. 39 Figure 4 - Map of Elkview property and sample site, located near Sparwood, southeastern, British Columbia. Tributaries that feed the Goddard Settling Pond drain directly off of the mine tailings of the open pit mine (DMGL, 1989). Location of Goddard Wetland with respect to open pit mine is circled and indicated by arrows. 40 Figure 5 - Layout of Goddard settling pond and sample sites in Goddard Wetland. Figure 6 - Photographs of a) decant of Goddard Settling Pond water into Goddard Wetland (note open pit mine in background with vegetation indicating mine tailings), and b) location of sample site #1 with respect to settling pond decant (indicated by arrow), a. b. SAMPLE SITE #1 42 Table 7 - Soluble sulphate and selenium concentrations in Goddard Settling Pond decant, monitored by Elkview Coal from April 1996 to April 2001. Date Dissolved Sulphate (mg/L) Total Selenium (ug/L) 04/16/1996 7.0 05/21/1996 - 7.0 08/13/1996 - 22.0 12/03/1996 - 16.0 03/04/1998 180.0 0.2 08/10/1998 193.0 0.4 12/02/1998 - 49.0 04/09/1999 58.9 -05/03/1999 61.9 -06/01/1999 98.0 -07/05/1999 51.7 -08/03/1999 107.0 -09/07/1999 132.0 -09/13/1999 305.0 91.0 10/04/1999 154.0 -11/01/1999 556.0 -12/06/1999 244.0 -01/03/2000 155.6 -02/07/2000 - 85.0 03/07/2000 219.2 -03/09/2000 229.5 -04/03/2000 92.9 -05/01/2000 182.0 -05/08/2000 174.0 41.0 06/05/2000 - 50.0 06/05/2000 187.7 -07/03/2000 - 88.0 07/03/2000 251.4 -08/01/2000 350.4 -08/29/2000 360.0 106.0 09/05/2000 250.0 -10/02/2000 324.8 -11/07/2000 - 93.0 11/07/2000 332.4 -12/05/2000 - 95.0 12/05/2000 509.4 -01/02/2001 - 91.0 01/02/2001 287.1 -02/06/2001 - 87.0 02/06/2001 364.5 -04/04/2001 - 55.0 43 3.1. MATERIALS AND METHODS 3.1.1. SAMPLE COLLECTION AND PRESERVATION On April 25, 2001, sediment samples were collected using a dredging or grabbing device built in the field. The dredge was composed of a 4-litre plastic water container tethered with wire to a 1.8-m aluminum pole (Figure 7). This type of sediment collecting apparatus was chosen based on sediment texture, depth of water, and presence of submerged vegetation. A random sampling method is usually used when quantitative data is required; however for the qualitative or semi-quantitative data desired for this study, a non-random sampling scheme was considered to be acceptable (Klemm et al., 1990). For the batch microcosm (Chapter 4) and adsorption experiments (Chapter 5), sample containers were 500-mL glass mason jars that had been previously sterilized for food preservation purposes. Approximately two dredges were required to fill each sample container for these studies. For the continuous microcosm experiment (Chapter 6), sediment samples were collected and stored in 4-L sealable plastic buckets. For all samples, each dredge was taken from the top 30-cm of sediment from an undisturbed area of the wetland. The dredge was rinsed off between samples with wetland water from an adjacent area to avoid contamination and disturbance of sample site. The sediment sample within each mason jar was topped off with water from the same site; this minimized the 44 presence of air bubbles and maintained the anaerobicity of the sample. All sediment samples in the study were handled identically, including using the same sampling apparatus, sampling method, sampling containers, and preservation techniques. Samples were preserved as soon as possible (within 36 hours of collection) by cooling to and maintaining a temperature of approximately 4°C (ice cold) in a large refrigerator, shaded from sunlight to prevent breakdown of chemicals by UV light. Figure 7 - This figure illustrates how the constructed dredge was used to sample adjacent site #5 in Goddard Wetland. We can see upstream of this pond that although there is vegetation below the open pit mine, these areas indicate mine tailings. 45 3.1.2. PRELIMINARY ANALYSIS OF SEDIMENT SAMPLES Concurrently, and in cooperation with my study, the B.C. Ministry of the Environment also collected samples from the Goddard Wetland. All water samples from the Goddard site were initially stored in air tight sterilized glass preserve jars as discussed in section 3.1.1. 248 mL of each water sample was transferred into a 250-mL sterile polyethylene bottle and preserved with 2 mL 1:1 nitric acid (samples Decant-H20 and S1-H20). Approximately 60 mL of sediment (from sample S1-A1) was transferred into a 120-mL sealable tissue culture cup. The sediment sample was not preserved with nitric acid. Preliminary analysis (using GF-AA) of these sediment and overlying water samples were performed by the Pacific Environmental Science Centre (North Vancouver). Water samples were analyzed for sulphate and total selenium, while sediment was only analyzed for total selenium. The results of these analyses are presented in Table 8. These results indicated that soluble Se in Goddard Pond decant was 28.2 ug/L and total soluble Se in Site #1 water averaged about 16.9 u.g/L. Total Se in Site #1 sediment was 39 ug/g dry weight. Total sulphate concentrations in Goddard Pond decant and Site #1 water ranged from 888 mg/L to 941 mg/L. 46 Table 8 - B.C. Ministry of the Environment preliminary analysis of Site #1 sediment and overlying water samples for Se and S; analysis performed by Pacific Environmental Science Centre in North Vancouver. Sample Type (collected by) Test Result Mean Detection Limit Goddard Pond Decant (MOE) SO?" 941 mg/L 26 mg/L Goddard Pond Decant (MOE) Se 28.2 ug/L 1 ug/L Goddard Site #1 water (UBC) SO?" 888 mg/L 26 mg/L Goddard Site #1 water (MOE) SO?" 896 mg/L 26 mg/L Goddard Site #1 water (UBC) Se 16.7 |ig/L 1 ug/L Goddard Site #1 water (MOE) Se 17.1 ug/L 1 ug/L Goddard Site #1 sediment (UBC) Se 39 ug/g (dry) 10 ug/g (dry) A carbon and nitrogen analysis was also performed on sediment samples from the Goddard Wetland Site (samples S1-A1 and S1-M1). Samples were completely mixed by repeatedly inverting sample jars for 1-2 minutes. At least 15-g (wet weight) of sediment was transferred into an acid washed 250 mL Pyrex Erlenmyer flask (refer to glassware preparation in Materials and Methods below). 30 mL of distilled, deionized water (hereafter referred to as dd-H20) was added to sediment in the flask to create a slurry, and then pH was measured. If the pH was found to be greater than 7, sufficient HCI was added to adjust pH below 7. Samples were oven dried over night and then finely ground using a pestle and mortar. The Leco CN-2000 Carbon and Nitrogen Analyzer was used to analyze the finely ground sediment samples (Application notes for this instrument can be 47 found in Appendix A). Synthetic carbon EDTA was used as a standard. The samples were combusted under oxygen at 1350°C; the carbon was determined by infrared absorption, and the nitrogen was determined by thermoconductivity. Results of this analysis are presented in Table 9. The percent of total carbon in sediment from sites # 1 - 5 ranged from 15.65% to 41.6%. Total nitrogen for each site ranged from 0.53% to 0.62%. Further analysis was performed on sediment to determine the percentage of oxidizable carbon (or non-coal carbon) in samples. This was done using a Walkley-Black Analysis (Allison, 1965). Sediment was prepared as discussed above for Leco CN-2000, after which 5 g of finely ground sediment was transferred into a 500 mL wide-mouthed Erlenmyer flask. 10 mL of 1 N K 2 Cr 2 0 7 was added to the flask and swirled gently to disperse the soil in the solution. In a fumehood, 20-mL of concentrated H 2 S0 4 was rapidly added by directing the stream of acid directly into the suspension. The flask was immediately swirled gently until soil and reagents were mixed, and then swirled more vigorously for exactly one minute. The flask cooled in the fumehood for 30 minutes after which 200 mL of dd-H 20 was added. This solution was titrated with 0.5 N FeS0 4. As the titration endpoint was approached, the solution took on a greenish cast and then changed to dark green. At this point, ferrous sulphate was added drop by drop until the colour changed from green to a very pale blue. A blank determination was performed in the same manner, but without our sediment sample, to standardize the Cr 20 72". Since more than 75% of the dichromate was 48 reduced, the determination was repeated with less soil (0.1 g dry weight). The results of this analysis (Table 9) were calculated according to the following formula, using a correction factor, / = 1.33. % Organic Carbon = (mEq K?Cr?Oz - mEq FeSQ4) (0.003) (100) Soil (g dry wt) The percentage of total oxidizable (or non-coal) carbon in sediment from all sites ranged from 2.39% to 16.36% (in Site #1); these results are included in Table 9. Table 9 - Total carbon, nitrogen, and total oxidizable carbon in surface sediments from Goddard Wetland. Sample Site % Total Carbon (+SD) % Total Nitrogen % Total Oxidizable Carbon S1 25.8 (± 0.99) 0.62 16.36 S2 18.35 (+ 0.78) 0.62 11.97 S3 15.65 (± 0.21) 0.57 9.58 S4 26.2 (± 0.95) 0.53 10.77 S5 41.6 (± 3.91) 0.62 2.39 3.2. SITE SURVEY AND ANALYSIS CONCLUSIONS 49 Following a thorough investigation of selenium mobilization from surface coal mining in the Elk River Basin, McDonald and Strosher (1998) determined that further investigation of selenium mobilization and accumulation in side channel wetlands in the Elk Valley was required. Preliminary analysis of Goddard Wetland Site 1 water by the Pacific Environmental Science Centre found that total Se in surface water averaged 16.9 ug/L; almost 17 times the current B.C. criterion (1 ug/L) for the protection of aquatic life (Nagpal et al., 1995). Further analysis of underlying sediment from this site measured a total Se concentration of 39 fig Se/g soil (dry weight). An analysis of sulphate in Site 1 water revealed an average concentration of 892 mg/L S042~. Due to high Se and So42" concentrations, as well as conditions at the site (pH 9.5, lentic system receiving settling pond drainage), sediment from this site was used for subsequent experiments. This site was appropriate for exploring the potential of bioremediatory treatment in wetlands, because of its slow turnover rate (greater retention time) and previous exposure to high Se and SO42" concentrations. Based upon high concentrations of sulphate and soluble selenium, growth media for sulphate-reducing bacteria and selenate-reducing bacteria were used to investigate the presence of these bacteria in our wetland sediment using batch microcosm experiments (Chapter 4). Incidentally, these two types of bacteria 50 share many characteristics, including the ability to reduce selenate to elemental selenium. Although not focused on in this study, high concentrations of sulphate in the Goddard Wetland might also pose a toxic threat to this system and therefore require treatment. Assessing the presence of SRB might also contribute to the mitigation of sulphate toxicity. Furthermore, this type of study can be used to determine whether any contaminated site and/or any contaminant can be remediated with bacteria that inhabit that area. In addition to batch microcosm experiments, an adsorption experiment was performed to determine the role sorption, plays in immobilizing soluble forms of selenium (Chapter 5). . Lastly, a continuous microcosm experiment was run to determine whether these selenium immobilizing/removal mechanisms would occur in an open system representative of the Goddard Wetland (Chapter 6). 51 4 . B A T C H M I C R O C O S M S T U D Y 4 . 1 . I N T R O D U C T I O N Batch microcosm studies were conducted three times following the collection of sediment from the Goddard Wetland. The initial batch microcosm study was conducted as pre-test material to assess the feasibility of selenate reduction under anaerobic conditions with two different bacterial growth media recipes. This pre-test provided information with regards to what duration batch microcosms should be allowed to run, as well as the identification of reaction products. A second preliminary study determined the optimal concentration of selenate that should be used to qualitatively examine changes in activity. The final batch microcosm was designed based on the results of preliminary tests to evaluate the potential for selenate reduction to elemental selenium (Se°). Selenate was used for this experiment because of its high mobility and lack of sorption in soils (refer to Chapter 5). The ultimate objective of this experiment was to determine whether anaerobic bacteria in Goddard wetland sediment are capable of immobilizing and/or removing the most mobile form of selenium in soil (selenate) via reduction. 52 4.2. MATERIALS AND METHODS For all batch microcosm trials, glassware and pipette tips used were thoroughly washed with soap, rinsed with tap water, soaked in 1% nitric acid (HN03) for 24 hours, rinsed with tap water four times, rinsed once with dd-H 20, and air-dried overnight. Batch microcosm protocol was adapted from similar methods used by Bradley et al. (1999) and Bradley and Chapelle (1997). Each microcosm treatment consisted of a sterilized 50-mL capped polypropylene Falcon tube that was amended with: a) sediment and growth media (Treatments 1-2), b) autoclaved sediment and growth media (Controls 3-4), c) only growth media (Controls 5-6), or d) inocula with respective growth media (Controls 7-8). Each treatment and control consisted of 4 replicates. These treatment descriptions can be found in Table 10. Controls 3 and 4 (consisting of autoclaved sediment and growth media) were included as killed controls. The only difference between these vials and those of Treatments 1 and 2 was the autoclaving of soil; this would indicate that changes observed in treatments were the result of microorganisms in vials. 53 Controls 5 and 6 (composed only of growth media) were included to ensure that contamination in vials was not responsible for changes observed in treatments. Controls 7 and 8 (growth medium and inoculum) were included in this experiment to conclude that our media was capable of growing microorganisms. Table 10 - Batch microcosm treatment descriptions. Treatment Sediment Inoculum Growth Media 1 10 mL OmL 40 mL SRB media 2 10 mL 0 mL 40 mL T.sel media 3 10 mL :autoclaved 0 mL 40 mL SRB media 4 10 mL :autoclaved 0 mL 40 mL T.sel media 5 OmL OmL 50 mL SRB media 6 0 mL 0 mL 50 mL T.sel media 7 0 mL 10 mL SRB inoculum 40 mL SRB media 8 0 mL 10 mL T.sel inoculum 40 mL T.sel media 54 4.2.1. MEDIA COMPOSITION SULPHATE REDUCING BACTERIA (SRB) POSTGATE MEDIUM C 1-L of Postgate C stock salt solution was made at ten times the desired concentration by adding: KH 2 P0 4 (5.0 g), NH4CI (10.0 g), Na 2 S0 4 (45.0 g), CaCI2-2H20 (0.40 g), MgS0 4-7H 20 (0.60 g), and FeS0 4-7H 20 (0.04 g) to 500 mL dd-H 20 in a 1 L Erlenmyer flask. Following addition of these salts the flask was topped up to 1 L with dd-H 20. A 1 L Erlenmyer flask holding 100 mL of stock salt solution (described above) was placed onto a stirrer at low speed and a clean stirbar added. 500 mL of dd-H 20 was added, followed by 4.5 mL of lactic acid, 0.3 g of sodium citrate-2H20, and 1 g of yeast extract. Flask is then filled to 980 mL mark with dd-H 20, and autoclaved for 30 minutes at 121°C and 15 psi. The flask is then removed from the autoclave and allowed to cool to room temperature. pH of the solution is raised to 7.5 by adding concentrated NaOH dropwise. It was required to wait several minutes between NaOH drops to allow the pH to stabilize. Following pH adjustment, 500 pg/L selenate was added to growth medium and flask was topped up to 1 L mark with dd-H 20. Selenate was added as a 500 u.g per 1 mL aliquot from a concentrated solution of sodium selenate (0.12 g Na 2Se0 4 in 100 mL dd-H20). Nitrogen gas is bubbled through the flask for 10 minutes, with the nitrogen source placed at the bottom of the flask in order for the 55 nitrogen to bubble up through the solution. SRB growth medium was then added to sediment samples as indicated in experimental method below. T. SELENATIS GROWTH MEDIUM Thauera selenatis media preparation techniques were adapted from Widdel and Pfenning (1984) and Oremland et al. (1994). The modified medium was prepared by adding the basal salts; K 2HP0 4 (0.225 g), KH 2P0 4 (0.225 g), NaCI (0.46 g), (NH 4) 2S0 4 (0.225 g), MgS0 4-7H 20 (0.117 g) and yeast extract (1.0 g) to 660 mL dd-H20 on a stirrer with a stir bar in an acid-washed 1 L Erlenmyer flask. This solution was then autoclaved for thirty minutes at 121 °C and 15 psi, and allowed to cool to room temperature. Following cooling, 1 mL of Trace Element Solution and 10 mL of each Vitamin Solution were added to the autoclaved solution. Trace Element Solution (Widdel and Pfenning, 1984) contained the following per liter of dd-H 20: HCI (6.58 mL of 25% solution), FeCI3 (1.5 g), CoCI2-6H20 (0.190 g), H 3 B0 3 (0.006 g), Na 3Mo0 4-2H 20 (0.036 g), CuCI2-2H20 (0.002 g). Vitamin Solution #1 was made per liter of 20 mM sodium phosphate buffer solution with pH 7.2. The following constituents were added to this buffer 56 solution: p-aminobenzoic acid (40 mg), D-biotin (15 mg), folic acid (40 mg), pyridoxamin-dihydrochloride (150 mg), niacinamide (100 mg), D-pantothenic acid (100 mg), thiotic acid (15 mg). Vitamin solution #2 contained riboflavin (25 mg) and thiamine chloride hydrochloride (100 mg) per liter of 25 mM sodium phosphate buffer solution (pH 3.7). Following addition of Trace Element and Vitamin solutions, 100-mL of lactic acid solution (1.7 mL lactic acid/100 mL dd-H20), and 100 mL of the electron acceptor selenate (3.78 g Na2SeO"4/100 mL dd-H20) are added to the Erlenmyer flask while gently stirring. Note that extreme caution should be taken when preparing and handling such a concentrated selenate solution. At this point, solution is deoxygenated by bubbling nitrogen gas through solution for ten minutes. Following deoxygenation, the remaining media ingredients are added to this solution in an anaerobic fumehood. These include 10-mL of each reducing agent (thioglycolic and ascorbic acids), as well as 100-mL bicarbonate solution (4.2 g bicarbonate/100 mL dd-H20). At this point, the growth medium should be topped up to exactly 1 L with dd-H 20 (note that a few millilitres of solution often evaporate off when solution has been autoclaved). The final pH of the growth media following bicarbonate addition and topping up should be approximately 7.3. T. selenatis growth medium was then added to sediment samples as indicated in experimental method below. 57 4.2.2. EXPERIMENTAL METHODS Sulphate reducing bacteria (SRB) growth media and T. selenatis (T.sel) growth media (discussed above) were prepared one day prior to incubation of treatments. Growth media were prepared identically for all three batch microcosm trials performed, except for one discrepancy; T.sel growth media for the second batch microcosm trial was prepared with 500 pg/L selenate rather than 3.78 g/L Na 2Se0 4. This trial will be discussed further in the Results section below. Following preparation of growth media, all batch microcosm procedures were performed in an anaerobic chamber. In an anaerobic chamber, contents of a previously sealed Goddard sample jar were poured into a larger plastic basin and homogenized thoroughly with plexiglass rods and shears. Mixing continued until all large matter (sticks, roots, etc.) were broken down or removed and sediment appeared to be a homogeneous slurry. Wetland sediment was then randomly sampled using a 10-ml_ volumetric pipette with a snipped pipette tip. For Treatments 1 and 2, 10 mL of sediment was dispensed into each of eight serum vials. Four of these serum vials (labeled treatment 1 a, b, c, and d) were then filled to brim with SRB growth media and sealed. Likewise, the other four serum vials (labeled treatment 2 a, b, c, and d) were filled to brim with T.sel growth media and sealed. 58 Treatments 3 and 4 served as killed controls using autoclaved sediment. 10 mL of sediment was dispensed into each of eight other serum vials, after which these vials were loosely capped, removed from anaerobic chamber, and autoclaved for 1 hr at 121°C and 15-psi. After cooling down to room temperature, four of these serum vials (labeled treatment 3 a, b, c, and d) were then filled to brim with SRB growth media and sealed. The other four serum vials (labeled treatment 4 a, b, c, and d) were filled to brim with T.sel growth media and sealed. Treatments 5 and 6 were meant to serve as sediment-free controls. This involved filling four serum vials to brim with SRB growth media (treatment 5 a, b, c, and d) and four serum vials to brim with T.sel growth media (treatment 6 a, b, c, and d). Rather than sediment, Treatments 7 and 8 were inoculated with 10 mL of SRB and T.sel inoculum, respectively. Inverting its bottle repeatedly for 1 minute adequately mixed each inoculum. Using an acid-washed, sterilized pipette, 10 mL of SRB inoculum was dispensed into each of four serum vials (labeled Treatment 7 a, b, c, and d). In the same manner, 10 mL of T.sel inoculum was dispensed into each of four serum vials (labeled Treatment 8 a, b, c, and d). Each set of 4 replicates was then filled to the brim with respective growth mediums and sealed. 59 All treatments were then placed in a 30°C incubator, and monitored and analyzed as discussed in the next section. 4.2.3. ANALYTICAL METHODS QUALITATIVE MONITORING All batch microcosm trials were monitored every 2-3 days using qualitative observations and/or digital photography. Any changes in colour of microcosm constituents and/or evolution of gas was recorded. Preliminary batch microcosm trial #1 (S1M1) was prepared following batch microcosm protocol discussed above. Incubation was begun on June 21-22, 2001 and continued until July 16, 2001. During this period, S1M1 microcosms were monitored with three sets of digital photographs (June 25, July 4, and July 10), found in Figure C1 (Appendix C). Preliminary batch microcosm #2 (S1M2) was not photographed, however qualitatively monitored throughout duration of incubation. 60 Treatment descriptions for all batch microcosm experiments (hereafter referred to as S1M1, S1M2, and S1M3) can be found in Table 10 in the Materials and Methods section above. SCANNING ELECTRON MICROSCOPE AND ENERGY-DISPERSIVE X-RAY ANALYSIS Red precipitates formed in S1M1 batch microcosms were observed under a Hitachi F-2300 Scanning Electron Microscope (SEM), and analyzed using Quartz-X1 Energy-Dispersive X-ray analytical equipment (EDX). For this analysis to be performed, sample had to be vacuum-compatible (i.e. not liquid or volatile) and electrically conductive. To prepare the sample for EDX, it was first dried under a heat lamp to ensure it was vacuum-compatible (i.e. not liquid or volatile). The sample was then made electrically conductive by coating it with a thin layer of carbon; the carbon did not interfere in any way with the analysis. EDX analysis looked at the x-rays produced in the sample by an electron beam. This analysis provided elemental information for the elements heavier than Boron on the periodic table. The detection limit was about 0.5 weight percent and the concentration was proportional to the area under the peak on the spectrum (refer to results). Analysis was performed using a 20 kV electron beam, for which 0 -20 keV x-ray energy was collected. 61 SULPHATE ANALYSIS Sulphate concentrations at the end of batch microcosm experiments 1 and 3 (S1M1 and S1M3, respectively) were determined with a Biochrom Ultrospec-1000 Spectrophotometer, using a standard turbidometric method (Greenberg et al., 1995). Rather than filtering samples prior to analysis, each microcosm was centrifuged for ten minutes at 3000 rpm. 9.9 mL of the supernatant was then transferred to a 15 mL sterile polypropylene Falcon tube with a volumetric pipette and preserved with 0.1 mL HN0 3. After turning on the spectrophotometer and allowing it to warm up for 20 minutes, a calibration curve was prepared. Using a standard sulphate solution and dd-H 20, six dilutions (with concentration 0, 10, 20, 30, 40, and 50 mg/L sulphate) were prepared. Absorbance readings for each of these dilutions were obtained using either Buffer A (for concentrations between 0 and 25 mg/L SO4 2"), or Buffer B (for concentration greater than 25 mg/L S0 42"). When making calibration curves, one was made for buffer A and one for buffer B. Standard absorbances, as well as sample absorbances were measured using the protocol outlined below. Standard calibration curves used for analysis of S1M1 and S1M3 trials can be found in Figure 8. 62 Using a pipette, 1 mL of sample was added to a small 100 mL beaker. Using another pipette, 19 mL of dd-H 20 was added to the same beaker. This diluted the sample twenty times. 4 mL of buffer was added to the beaker. A clean small stir bar was added to the beaker, and beaker was placed on a stirrer (set to low speed). All samples used exactly the same stir speed. Half a gram of barium chloride powder was weighed out into a plastic weigh boat. This was added to the beaker while starting a stopwatch at the same time. The beaker is stirred with this excess of barium chloride for exactly 60 seconds, after which the solution is poured into a cuvette and let to sit for an additional five minutes. At exactly five minutes, the absorbance reading was measured on the spectrophotometer. Sulphate concentration of each batch microcosm sample is determined by comparing absorbance readings with those of the standard curve. 63 Figure 8 - Standard calibration curves used for sulphate analysis of a) S1M1 and b) S1M3 trials. 0.3 -i 0.25 0.2 -o 0.15 H < 0.1 0.05 a. S1M1 SO4 STANDARD CURVE y = 0.0077X- 0.1357 y = 0.0042x + 0.0255 10 20 30 40 Concentration (ppm) Buffer A Buffer B 50 60 b. S1M3 S 0 42 " STANDARD CURVE 60 80 Concentration 120 64 SELENIUM ANALYSIS (HYDRIDE GENERATION-ATOMIC FLUORESCENCE) At the end of experiments, each batch microcosm Falcon tube was centrifuged for ten minutes at 3000 rpm, after which 9.9 mL of the supernatant was transferred to a clean 15 mL Falcon tube, and preserved with 0.1 mL HN0 3 (1% acidification as performed for sulphate analysis above). Samples were stored at 4°C until analysis. All glass- and plastic-ware utilized during Se analysis (centrifuge tubes, pipette tips, marbles, beakers, autosampler vials) were thoroughly washed using the following procedure: i) Rinsed with tap water twice; ii) Soaked in 5% nitric acid bath for 24 hours; iii) Rinsed with tap water once; iv) Soaked in 5% HCI bath for 24 hours; v) Rinsed with dd-H 20, and vi) Air-dried overnight. For analysis, selenate had to first be reduced to selenite using the following method. A 1 mL aliquot of sample was transferred to a 100 mL tared glass test tube. Sample was dispensed by volume, however weight was also recorded to ensure accurate dilution of sample. 20 mL of high purity hydrochloric acid was added to sample in centrifuge tube, followed by 10 mL of dd-H20. A 65 glass marble was placed on top of the tube, and sample was then boiled gently for 30 minutes at 135°C in a block digester. After samples were cooled to room temperature, 19 mL of dd H 2 0 was added to the sample. Each test tube (plus contents) was weighed in between each step to ensure accuracy and determine vapor loss during digestion, which was minimal. This acid digestion resulted in a matrix of 40% HCI (identical to reagent and standard blanks). Five selenium (IV) calibration standards (0, 0.1, 1,5, 10 pg/L) were prepared by serial dilution of a commercial quantitative selenium (IV) standard (1000 mg/L). As previously mentioned, each standard is prepared in a 40% HCI matrix. Total selenium concentration for each aqueous sample was determined using hydride generation-atomic fluorescence (HG-AF) on a Millenium Excalibur PSA 10.055. This system consists of an autosampler, an integrated continuous flow vapour generator, an atomic fluorescence spectrometer, and control computer. Acidified sample solutions were treated with sodium tetrahydroborate to generate the covalent gaseous hydride (selenium hydride). The hydride and excess hydrogen were swept out of the generation vessel using a stream of argon, into a chemical generated hydrogen diffusion flame. The hydrides were atomized and the resulting atoms detected by atomic fluorescence spectrometry 66 (Corns et al., 1993). The Millenium Excalibur method for selenium in drinking, surface, ground, saline, industrial, and domestic waters was automated by means of an autosampler and Avalon™ control software. On the continuous flow system, a reagent blank solution is run as background for automatic blank subtraction. Each 1000 mL of reagent blank solution contains 400 mL hydrochloric acid. This solution may contain trace levels of detectable amounts of selenium, therefore it is important that the same reagents are used for sample, standard, and reagent blank preparation. 0.7% m/v sodium tetrahydroborate (NaBH4) in 0.1 mol/L sodium hydroxide (NaOH) is prepared by dissolving 7.0-g of NaBH4 in 500 mL dd-H 20 and adding 4.0-g NaOH. This solution is then diluted to 1000 mL with dd-H20. This solution should be prepared freshly each day, and not kept in a closed container because of pressure build-up due to hydrogen evolution. Note extreme caution was practiced when working with NaBH4; it is extremely toxic and reacts violently with acids. Instrumentation of this equipment should be configured as shown in Figure 9 and described in the Millenium Excalibur User Manual (1997). 67 Figure 9 - Schematic configuration of Atomic Fluorescence System - Millenium Excalibur PSA 10.055 (Millenium Excalibur User Manual, 1997). BLANK SAMPLE RECYCLE REDUCTANT GAS/LIQUID SEPARATOR 68 4.3. RESULTS 4.3.1. QUALITATIVE ANALYSIS The appearance (mainly colour) of batch microcosms was monitored and recorded using a digital camera. The first batch microcosm trial (S1M1) experienced some problems; results from this trial indicated that autoclaving was not long enough, the SRB inoculum control did not work, and monitoring was not performed frequently enough. The second batch microcosm trial (S1M2) used a much lower Se concentration media, and thus no red precipitate was observed. These results were taken into consideration for the third batch microcosm trial (S1M3). Selected images for S1M3 are shown in Figure 10. TREATMENT 1 - SEDIMENT AND SRB MEDIUM The purpose of this treatment was to test for the presence of sulphate-reducing bacteria (SRB). In all three trials, no.changes were observed until day 12 or 13, when blackening of the solution occurred (due to the formation of FeS) and gas-bubbles (H2S) were observed. No red precipitate, indicative of Se°, was seen. 69 TREATMENT 2 - SEDIMENT AND T. SEL MEDIUM The purpose of this treatment was to test for the presence of selenate-reducing bacteria (SeRB). In trials run with high Se concentration media, a light pink colour was observed one day after inoculation. Within 4 days after inoculation, a dark red precipitate could be seen on Falcon tube walls and in the upper layers of sediment. For the trial performed with low Se concentration media (S1M2), no pink or red colour appeared. TREATMENT 3 - AUTOCLAVED SEDIMENT AND SRB MEDIUM CONTROL This control tested for any non-microbial interaction between SRB growth medium and sediment. In the first trial (S1M1), blackening occurred and thus it was postulated that autoclaving had not killed all SRB in the sediment. Autoclaving time was doubled for subsequent trials (S1M2 and S1M3). Still, a slight blackening did start at the end of these trials, however much less than for unautoclaved treatments. 70 TREATMENT 4 - AUTOCLAVED SEDIMENT AND T. SEL MEDIUM CONTROL This control tested for any non-microbial interaction between SeRB growth medium and sediment. Similarly to Treatment 3, due to insufficient autoclaving time, trial S1M1 exhibited a red colour. Subsequent trials were autoclaved longer, and remained unchanged until the very end of monitoring period (14 days), when a slight pink colour appeared in the Falcon tubes. TREATMENT 5 - SRB MEDIUM CONTROL For all batch microcosm trials, this control (containing only SRB medium) remained clear throughout the duration of each experiment. TREATMENT 6 - T. SEL MEDIUM For almost all batch microcosm trials, this control (containing only T. selenatis medium) remained clear throughout the duration of each experiment. Only one vial (S1M1 Treatment 6a) exhibited slight pink haze with minimal formation of red precipitate; contamination may have occurred here. 71 TREATMENT 7 - SRB INOCULUM AND SRB MEDIUM CONTROL This control shows that the prepared growth medium can actually support SRB. Seven days after inoculation, the solution became much cloudier (darker) in colour. 12 days after the start of this experiment solution became jet black, and a prominent evolution of gas was observed (as bubbles in solution and pressure build-up in vials). TREATMENT 8 - SeRB INOCULUM AND T. SEL MEDIUM CONTROL This control shows that the prepared growth medium can support SeRB. In S1M3, red precipitate appeared almost immediately after inoculation (between days one and two). Three to six days after the start of this experiment, the red precipitate in all replicates became most intense (bright red). Between days seven and eight, the red precipitate became much darker in colour, and by day 12 solution had become transparent again with red fading to an orange-brown tint. Figure 10 - Digital photograph monitoring of batch microcosm trial S1M3. 1 6 / 1 1 /2001 T1 - T2 T3 - T4 T5 - T6 T7 - T8 1 9 / 1 1 /2001 T1 - T2 T3 - T4 T5 - T6 T7 - T8 2 0 / 11 /2001 T1 - T2 T3 - T4 T5 - T6 T7 - T8 Figure 10 Continued... 21 / 11 /2001 M l s T1 - T 2 22 / 11 / 200' T3 - T4 T5 - T6 T7 - T8 w w P ^ 99 II i i PH X T ¥ T I T1 - T 2 27/11 /2001 T3 - T4 T5 - T6 T 7 - T 8 w T1 - T2 T3 - T4 T5 - T6 T7 - T8 Footnote: For all weeks in Figure 11 above, tubes 2 and 4 have been mislabeled Figure 10 Continued. 21 /11 /2001 T1 - T2 T3 - T4 T5 - T6 T7 - T8 22 / 11 / 2001 MM pW MB IN T1 - T2 T3 - T4 T5 - T6 T7 - T8 27/11 /2001 T1 - T2 T3 - T4 T5 - T6 T7 - T8 Footnote: For all weeks in Figure 11 above, tubes 2 and 4 have been mislabeled 74 4.3.2 SCANNING ELECTRON MICROSCOPE (SEM) AND ENERGY-DISPERSIVE X-RAY ANALYSIS (EDX) This analysis was performed with red precipitate obtained from batch microcosm S1M1 treatment 2d (sediment + T.sel medium), so as to confirm the presence of elemental selenium. Scrapings of red precipitate taken from walls of sample 2d were photographed under scanning electron microscope (SEM) at 600 times magnification (Figure 11). Constituents of these scrapings were analyzed using energy-dispersive x-ray (EDX) analysis. The first scraping contained approximately 24% (dry weight) elemental selenium, and the second scraping contained approximately 40% (dry weight) elemental selenium. Other forms of selenium (i.e. FeSe2 and Se-oxides) were excluded from analysis. A third scraping was also analyzed from the surface of the sediment in the same sample vial; this scraping contained only 3.46% (dry weight) elemental Se. These results can be found in Tables 11 a-c. 75 Figure 1 1 - Scrapings of red precipitate taken from walls of sample 2d; photographed under scanning electron microscope (SEM) at 600 times magnification. 76 Tables 11 a,b,c - Energy-dispersive x-ray (EDX) analysis for a) scraping #1 from wall of serum vial, b) scraping #2 from wall of serum vial, and c) scraping #3 from surface of sediment, a) Element Concentration (by % weight) Oxygen 33.5 % Aluminum 8.49 % Silicon 22.77 % Sulfur 0.53 % Potassium 2.18% Calcium 5.01 % Titanium 0.52 % Iron 2.94 % Selenium 24.08 % ) Element Concentration (by % weight) Oxygen 25.43 % Aluminum 12.57 % Silicon 12.51 % Sulfur 1.02 % Potassium 0.85 % Calcium 6.42 % Iron 1.11 % Selenium 40.10 % ) Element Concentration (by % weight) Oxygen 42.58 % Aluminum 12.66 % Silicon 27.49 % Sulfur 0.35 % Potassium 6.07 % Calcium 2.86 % Titanium 0.36 % Iron 4.17% Selenium 3.46 % 77 4.3.3 SULPHATE ANALYSIS Sulphate analyses were performed on S1M3 treatments containing SRB media. Sulphate reduction further confirmed the presence of SRB in Goddard wetland sediment. In the SRB growth media control (Treatment 5) where no sediment or inoculum was added, SO42" concentrations ranged from 397 ppm to 410 ppm with an average concentration of 408 ppm ±11 ppm. Treatment 1 S0 42 " concentrations (sediment + SRB media, N=4) ranged from 152 ppm to 167 ppm with an average concentration of 157 ppm ± 6.9 ppm. Treatment 3 SO42" concentrations (autoclaved sediment + media, N=3) ranged from 182 ppm to 258 ppm with an average concentration of 222 ppm ± 38 ppm. Lastly, treatment 7 SO42" concentrations (N=4) ranged from 261 ppm to 359 ppm with an average concentration of 315 ppm ± 51 ppm (Figure 12). 78 Figure 12 - Summary of sulphate analysis for batch microcosm S1M3. 79 4.3.4 SELENIUM ANALYSIS All treatments that used the SRB medium (Treatments 1, 3, 5, and 7) were prepared with a theoretical concentration of 500-ppb selenate (as Na2Se04). Measured Se concentrations from these treatments are presented in Figure 13. Analysis of Treatment 5 vials (containing only SRB medium), indicated that the average concentration (N=4) in our SRB medium was actually 619.0 ± 52.7 ppb total Se. Treatment 1 (composed of SRB medium and sediment) was found to have an average total Se concentration (N=3) of 15.1 ± 0.65 ppb. This was approximately 40 times less selenium than the control containing only growth medium. Treatment 3 was a killed control (containing SRB medium and autoclaved sediment). The average total selenium concentration for 4 replicates was 78.5 ± 20.59 ppb, almost eight times less selenium than control Treatment 5. Treatment 7 (composed of SRB inoculum and SRB medium) had an average total Se concentration of 21.5 ± 7.6 ppb (N=3), 28 times less than Treatment 5 control. 80 Figure 13 - Selenium concentrations measured in S1M3 treatments containing SRB medium. 1 3 5 Treatment 7 81 All treatments amended with T. selenatis growth medium had a theoretical selenium concentration of 1.6 g/L selenate (or 1580 ppm, added as 3.8 g Na2Se0 4 / L medium); however upon analysis of the control Treatment 6 (T.sel medium only), the average total Se concentration (N=4) was determined to be 364 ± 66 ppb. Treatment 2 consisting of T.sel medium and sediment) had an average total Se concentration of 22 ± 10 ppb based upon 3 replicates. This Se average was approximately 16.6 times less than the average for control Treatment 6, and almost 13 times less than the average for killed control Treatment 4. Treatment 2, was a killed control; T.sel medium with autoclaved sediment. The average Se concentration for this treatment (N=4) was 285 ± 67 ppb. This value falls within the range of control Treatment 6. Treatment 8, composed of selenate-reducing bacteria inoculum and T.sel medium, had an average concentration of 2.5 ± 0.2 ppb total Se. This was the lowest Se concentration of all treatments, with an average almost 150 times less than the average concentration for control Treatment 6 (Figure 14). 82 Figure 14 - Selenium concentrations measured in S1M3 treatments containing T.sel medium. 83 4.4. DISCUSSION Three batch microcosm trials (S1M1, S1M2, and S1M3) were performed for this study. All trials experienced some experimental difficulties, however these results did provide useful information for running subsequent trials. This discussion will outline these experimental difficulties, the rationale for particular methods used, and significance of the results. The first batch microcosm trial (S1M1) determined whether any qualitative data could be obtained from this research, including colour changes of solution, formation of precipitate, and evolution of gas. S1M1 also indicated for what duration microcosms should be monitored. Treatments amended with T. selenatis growth medium indicated that the majority of qualitative changes (namely formation of red precipitate) occurred within two weeks of inoculation. Qualitative changes in treatments using SRB growth medium did not occur as quickly, therefore these treatments were monitored for approximately 3 weeks. Killed controls (using autoclaved sediment) were prepared to ensure that any variance between qualitative results of treatments and killed controls could be attributed to sediment microorganisms and not other confounding factors. This was proven effective based on the results of trial S1M3; treatments 1 and 2 (sediment +medium) both exhibited qualitative changes in colour and/or evolution of gas, while treatments 3 and 4 remained clear and exhibited minimal colour change throughout duration of experiment. Quantitative Se analysis of these treatments (refer to results) confirm these results. Note that all trials eventually exhibited some qualitative changes in killed controls; S1M2 and S1M3 killed controls were autoclaved longer (1 hour), however both displayed some growth towards the end of the experiment. For future studies, much longer autoclaving periods are required to ensure totally sterile sediment. Sulphate concentrations in controls containing only SRB growth media (Treatment 5) were much lower (397 ppm - 410 ppm) than the theoretical sulphate concentration expected in SRB medium (3067 ppm). This was probably the result of a dilution error; this however did not impact the integrity of results since there was still significant growth in the SRB inoculum control (Treatment 7). This was further substantiated by a significant difference in sulphate concentration between the SRB media-only control (Treatment 5) and SRB medium inoculated with sediment (Treatment 1). S1M1 treatments provided identification of a red precipitate as elemental selenium. Oremland et al. (1994) isolated a selenate-reducing bacteria (T. selenatis - strain SES-3) from an acetate-oxidizing, selenate-respiring enrichment recovered from a freshwater marsh. The enrichment was streaked on 2% agar plates composed of a mineral salts medium (20 mM acetate plus 20 mM selenate). After being stored anaerobically for a few weeks, red colonies were 85 isolated from the agar plates. These colonies were transferred to culture tubes with liquid medium and sealed anaerobically. Growth was evident as turbidity in the tubes and formation of red elemental Se°. The medium used in our research was an adaptation of the recipe prepared by Oremland et al. (1994). Based upon similar results in batch microcosm S1M1, the observance of red precipitate was considered to be a positive identification of selenate reduction to elemental Se in future batch microcosms (S1M2 and S1M3). The measured selenate concentration in the T.sel media-only control (Treatment 6) was much less (364 ppb) than the theoretical amount of selenate added to prepare T.sel media (1580 ppm). This may be explained by human error (such as dilution error or insufficient mixing). However, this did not negatively impact the results of this study, since the reduction of selenate to elemental Se° was still observed (as the formation of a red precipitate). Batch microcosm trial S1M2 was performed with a lower concentration of selenate. This determined whether a higher (3.78 g/L) or lower (500 pg/L) concentration of selenate would result in more apparent qualitative changes in batch microcosms. The lack of qualitative results for S1M2 T.sel treatments as compared to the observance of red precipitate formation in S1M1 and S1M3 may have some important implications for this type of remediation process. Qualitative results may indicate that reduction of selenate to elemental Se is most effective in treating extremely high concentrations of selenate. However, 86 this conclusion can not be justified without further quantitative analysis of batch microcosms using lower concentrations of selenate. Distinctions between reduction of selenate by Se-reducing bacteria and SRB are discussed later in this section. The evolution of gas was monitored in all microcosm trials as bubbles occurring in solution or build-up of pressure within each serum vial. Although headspaces were not analyzed to determine products/by-products of reactions occurring in each serum vial, these gases could potentially include volatile selenium compounds (i.e. dimethylselenide) or products of sulphate reduction (H2S). For future batch microcosm studies, it is recommended that analysis of headspace be included in analytical methods. All batch microcosm trials were prepared and sealed in an anaerobic chamber to maintain anaerobicity of the sediment sample. Samples were kept anaerobic to mimic suspected anaerobic conditions existing in Goddard sediment. Although redox potential was not measured at the site, we postulated that very low flow (long retention time) in sample site 1 contributed to anerobic conditions. For future studies, redox potential should be measured in the field. 87 SULPHATE ANALYSIS According to the recipe for Postgate C stock salt solution (Pg.54), the sulphate concentration in SRB growth medium should have been 3067 mg/L. However, the measured sulphate concentration in the growth medium control (Figure 12) was 408 mg/L. As quantitative changes indicative of S0 42 " reduction were not observed (blackening of solution), this discrepancy may be due to an error in preparation of this medium Sulphate concentrations for S1M1 treatment 1 (media and sediment) averaged 828 ppm ± 42 ppm, while treatment 3 (media and autoclaved sediment) sulphate concentrations averaged 922 ppm ± 34 ppm S0 42 " (refer to Figure D1 in Appendix D). A single-factor analysis of variance (ANOVA) was performed to test the hypothesis that the means from treatment 1 and treatment 3 are statistically different. The source of variation between these two treatment groups has a P-value of 0.03, therefore we can accept the hypothesis that these treatment groups are significantly different. The same analysis of variance performed for Trial S1M3 found that treatments 1 and 3 were statistically different (P-value = 0.02), with treatment 3 having an average sulphate concentration almost 1.5 times higher than treatment 1. Although treatment 3 was autoclaved longer (1 hour), some reduction of sulphate was still observed. 88 S1M3 treatment 1 (sediment and SRB medium) and treatment 5 (SRB medium) were also found to be statistically significant with an obvious reduction of sulphate by treatment 1 to almost 1/3 of media-only control (P-value -> 0.00). Treatments 5 and 7 were also statistically different (P-value = 0.011), however a significant reduction of sulphate was not observed. This may be attributed to the use of an old inoculum in treatment 7 in which most of the SRB had already died off. S1M3 SELENIUM ANALYSIS For S1M3, SRB medium was prepared with a theoretical concentration of 500 ppb selenate (as Na2Se04), however the average concentration of SRB medium (Treatment 5) was actually 619 ppb ± 53 ppb total Se. This may be attributed to human error in preparation of medium. S1M3 Treatment 1 (SRB medium and sediment) was found to be very effective in reducing selenate in SRB medium (619 ppb + 53 ppb total Se) to almost 1/40 of this concentration (15 ppb ± 0.7 ppb). Analysis of variance confirms that treatment 1 and 5 are statistically different (P -> 0). Treatment 3 (killed control) had an average total selenium concentration of 79 ± 21 ppb, almost eight times less selenium than control Treatment 5. 89 Although it was autoclaved, perhaps some SRB did persist to reduce Se(VI) to Se(0), however this killed control was still found to be significantly higher than Treatment 1 (P-value = 0.003). Treatment 7 (composed of SRB inoculum and SRB medium) had an average total Se concentration of 22 ± 8 ppb (N=3), 28 times less than Treatment 5 control. This treatment was found to be statistically similar to treatment 1, indicating that bacteria in Goddard sediment were as effective (if not more effective) in reducing Se(VI) to Se(0) as an established SRB inoculum. T. selenatis growth medium was prepared with a theoretical selenium concentration of 1.6 g/L selenate (or 1580 ppm, added as 3.8 g Na2Se04/ L medium ). Upon analysis of control Treatment 6 (T.sel medium only), the average total Se concentration was determined to be 364 ppb ± 66 ppb; this value was much lower than anticipated. This discrepancy might be attributed to a dilution error, insufficient mixing of solution, or perhaps an inaccurate measurement of sodium selenate powder. Regardless, the average concentration of 364 ppb measured in treatment 6 vials was considered the original media Se concentration for all treatment vials. Treatment 4 (killed control) had an average Se concentration of 285 ppb ± 67 ppb. This value was not found to be significantly different (P = 0.14) from 90 control Treatment 6 (T.sel media-only). Therefore, autoclaving treatment 2 sediment longer did help to kill off SeRB. Treatment 2 (T.sel medium and sediment) had an Se average that was approximately 16.6 times less than the average for control Treatment 6, and almost 13 times less than the average for killed control Treatment 2. However, treatment 8 (established SeRB inoculum and T.sel medium) exhibited the greatest reduction of selenate, with an average almost 150 times less than the average concentration for control Treatment 6. This would indicate that although SeRB exist in our sediment, they are not as proficient at reducing selenate to elemental Se as the established SeRB inoculum used in Treatment 8. Batch microcosm trial S1M3 was not analyzed for Se until several weeks (45 days) after the end of incubation period. Although bright red and black precipitates may have faded in S1M3 treatments, a significant reduction of soluble Se was measured in treatment serum vials. This justifies that precipitates formed are stable in reducing environments, insoluble in water, and oxidize very slowly (Haygarth, 1994). 91 4 . 5 . BATCH MICROCOSM CONCLUSIONS AND RECOMMENDATIONS Goddard Wetland sediment tested positive for the presence of selenate-reducing bacteria, and SRB, which were shown to also reduce selenate. This suggests that microbiological reduction of selenate in Goddard Wetland is possible. Although these trials indicate a potential for using these microorganisms to reduce aqueous selenium concentrations, future studies should be performed to follow up on these results. Experiments should be run to determine the actual presence and activity of bacteria in Goddard Wetland sediment. The abundance of these microbial populations in sediment samples could be determined by analyzing sediment samples with a Most-Probable-Number (or dilution) method. This study would estimate the size of microbial populations based on the highest dilution at which growth could be obtained. The activity (or capacity) of these microorganisms to reduce Se might best be determined by repeating similar batch microcosm studies. These studies could include running several identical microcosms, and analyzing each one at different times after inoculation to determine the rate of Se reduction. Also, an investigation of SRB/SeRB growth kinetics could be performed to show how the growth of these microorganisms is influenced by the presence of Se oxyanions. 92 In future studies, enrichment cultures could be prepared from batch microcosms that exhibited growth of SeRB. An aliquot of culture solution from the batch microcosm could be used to inoculate fresh T.sel media and then incubated. Upon observing the proliferation of SeRB in this first enrichment culture, this process could be repeated with transfer to more fresh media. These enrichment cultures could be used to optimize the growth of Se-reducing bacteria (and subsequent reduction of Se). Ultimately, these enrichment cultures could be put to use in bioreactors designed to treat Se-laden effluent. The formation of red precipitate (indicative of Se reduction) was not observed when sediment and/or inoculum was amended with T.sel media made with 500-p.g/1 Se (trial S1M2). Although qualitative changes did not occur in these microcosms, for future studies a quantitative Se analysis (HG-AF) should be performed to assess whether significant reduction of Se occurred with SeRB and lower concentration of selenate. The formation of a red precipitate was not observed in any SRB batch microcosm treatments (treatment 1 and 7), however these treatments did change in colour (becoming black and cloudy). Se analysis of trial S1M3 indicated that treatments 1 and 7 did exhibit a marked reduction in soluble Se (refer to HG-AF analysis results). 93 SeRB and SRB could be used in a treatment system consisting of a mesocosm, or surface flow constructed wetland. This system could be situated to collect and treat Se-contaminated waters prior to their entering a natural wetland, and also serve to simulate natural wetland conditions. Furthermore, the rate of selenate reduction to elemental selenium in this mesocosm could be optimized by testing various applications. This may involve applying specific growth media or organic amendments to induce the growth or activity of microbiota endogenous to wetland sediment (also known as biostimulation). In conclusion, this study indicates that the full potential of biotransformative Se reduction in wetland sediment and remediation technologies is yet to be determined. 94 5. ADSORPTION STUDY 5.1. INTRODUCTION The batch microcosm study (Chapter 4) indicated that SRB and SeRB were present in Goddard Wetland sediment, and that reduction of selenate could potentially occur in this wetland. However, adsorption of selenium oxyanions to particulates or biomass in the sediment could also contribute to reducing concentrations of Se in solution. In this Chapter, I outline and discuss an adsorption study performed to determine the potential and extent of Se adsorption by Goddard Wetland sediment. The ultimate objective of this study is to assess whether Goddard wetland sediment is capable of adsorbing the mobile forms of selenium in soil (selenite and selenate) and thereby acting as a sink for selenium. 5.2. MATERIALS AND METHODS Glassware and pipette tips used for the adsorption study were thoroughly washed with soap, rinsed in tap water, soaked in 1% nitric acid (HN0 3 ) for 24 hours, rinsed with tap water four times, rinsed once with dd-H20, and air-dried overnight. 95 Adsorption study protocol was adapted from similar methods used by Guo et al. (1999 a and b). Adsorption treatments consisted of sterilized 15 mL capped polypropylene Falcon tubes that contained sediment and selenium concentrations (as sodium selenite or sodium selenate) ranging from 0 ppb to 500 ppb (Table 12). Sediment sample S1A1 used in this study was collected and preserved as previously described for batch microcosm sediment in Chapter 4. This study was performed twice; once with sodium selenite and once with sodium selenate. A more thorough description of various treatments and controls is discussed in the following section. 5.2.1. EXPERIMENTAL METHODS All glassware and instruments used (except pre-sterilized one-use Falcon tubes) were washed following procedure discussed above. Sample jar S1A1 was autoclaved for one hour at 121°C and 15 psi. During autoclaving, the sample jar was covered with a tinfoil lid to minimize evaporation of moisture in the sample. The autoclaved sample was then transferred to a generic food-processing blender, and homogenized thoroughly for 10 minutes. The homogenized sample was dried overnight in an oven at 90°C to remove any moisture from homogenous slurry. The dried sample was 96 removed from the oven and re-homogenized using a pestle and mortar until a fine homogenous mixture was observed. Two gram aliquots of sample (dry-weight) were placed into each of 40 15 mL capped, polypropylene Falcon tubes. In addition to these 40 tubes, 12 Falcon tubes were prepared as controls without soil. Based on methods of soil analysis by Huang and Fujii (1996), adequate extractants used to wash soil of any pre-existing adsorbed selenite and/or selenate are KCI and KH 2P0 4; therefore, each tube was filled to the brim with a 0.25 M NaCI - 0.1 M KH 2P0 4 deoxygenated solution and sealed. All tubes (including controls) were sealed, laid horizontally in a box, and shaken vigorously for 72 hours. After 72-hours of shaking, all tubes were centrifuged (15-minutes at 3000 rpm), and the supernatant was decanted from each tube. Each tube was then refilled to brim with deoxygenated, distilled water, and agitated for 30 seconds using a vortex mixer. Following this agitation, all tubes were re-centrifuged and decanted as before. This centrifuging / washing procedure was repeated four times to ensure adequate washing of soil (removal of soluble Se species and extractants, NaCI and KH 2P0 4). 97 After centrifuging and decanting water on fourth wash, 10-mL of appropriate de-oxygenated Se solution was added to each tube (refer to schematic Figure 15) and all tubes were adequately sealed. All measures were taken to minimize oxidation of sediment sample S1A1; this included de-oxygenating all solutions for 10-minutes with N g a s and purging all headspaces with N g a s prior to sealing tubes. Also, homogenization times and exposure of sediment to air was kept to a minimum. After addition of Se solutions, all tubes were shaken for 48 hours (as before) and centrifuged for 15 minutes at 3000 rpm. The supernatant from each tube was then decanted into a pressure filter (through 0.22 pm filter paper). For each sample, two 4 mL aliquots of filtered solution were transferred from the filter trap into two small, labeled, capped culture tubes (5 mL capacity, sterilized polypropylene), and each tube preserved with 40 pL Environmental grade HNO3 (final preserved concentration = 1%). 98 Table 12 - Concentrations of Se corresponding to concentrations of sodium selenite (Na2SeC>3) and sodium selenate (Na2Se04). [Na2Se03] = [SelV] [Na2Se04] = [SeVI] 500 228.3 500 209.0 250 114.2 250 104.5 125 57.1 125 52.2 62.5 28.5 62.5 26.1 31.25 14.3 31.25 13.1 15.625 7.1 15.625 6.5 Figure 15 - Schematic of adsorption experiment run on sample sediment S1A1. with soil T w/out soil (controls) sodium selenite sodium selenate O O O o © © o o o o o © o o o o o © o o o © © © o o o © © © o o o 0 0 0 0 0 0 0 0 0 0 0 0 0 0 © 0 0 0 0 500 ppb (a-c) 250 ppb (a-c) 125 ppb (a-c) 62.5 ppb (a-c) 31.25 ppb (a-c) 15.625 ppb (a-c) 0 ppb (controls) 500 ,250, and 125 ppb 62.5,31.25, and 15.625 ppb 100 5.2.2. ANALYTICAL METHODS SELENIUM ANALYSIS (HG-AF) Total selenium concentration was determined for solution from each adsorption trial. This analysis was performed using the Atomic Fluorescence System - Millenium Excalibur PSA 10.055. An in-depth description of this equipment, the analytical methods used to prepare samples, and the determination of total soluble Se using HG-AF can be found in Chapter 4.2.3. 5.3. RESULTS 5.3.1. HYDRIDE GENERATION-ATOMIC FLUORESCENCE (HG-AF) SELENIUM ANALYSIS Results of selenite and selenate adsorption trials are provided in Table 13 and Table 14, respectively. Theoretical Se concentrations were calculated based on the addition of 500, 250, 125, 62.5, 31.25, and 15.625 ug/L of sodium selenite or sodium selenate. 101 In determining the final concentration of selenite or selenate, a background concentration was subtracted from all treatments. Treatments that were supposed to have a theoretical Se concentration of zero (consisting only of deoxygenated, distilled water), ended up with some selenite or selenate being picked up during analysis. This may be the result of some Se leaching out of sediment, therefore this value was subtracted from final Se concentrations. Table 13 - Selenite Adsorption Trial Results (difference between [Se] t r eatment and ave [Se]treatment due to subtraction of background value, discussed above). [Na 2Se0 3] (PPb) [Se]theoretical (PPb) [Se]Control (PPb) [Ssjtreatment (PPb) ave [Se]treatment (PPb) SD 500 228.3 222.5 16.9 5.3 4.1 250 114.2 76.7 19.6 8.0 3.3 125 57.1 42.6 13.7 2.1 2.5 62.5 28.5 33.7 18.7 7.1 8.4 31.25 14.3 18.9 15.0 3.4 4.6 15.625 7.1 8.4 7.9 -3.8 0.0 1.5 0 0.0 - 11.6 = background 0.0 2.6 1 0 2 Table 14 - Selenate Adsorption Trial Results (difference between [Se] t r eatment and ave [Se]treatment due to subtraction of background value, discussed above). [Na 2Se0 4 ] (PPb) [Se]theoretical (PPb) [Se]Control (PPb) [Se]treatment (PPb) ave [Se]treatment (PPb) SD 500 209.0 125.9 151.9 145.7 23.0 250 104.5 92.8 83.7 77.6 9.3 125 52.2 27.7 44.5 38.4 5.7 62.5 26.1 12.6 22.5 16.4 3.5 31.25 13.1 8.0 18.9 12.8 2.3 15.625 6.5 8.0 14.5 8.4 3.2 0 0.0 - 6.1 = background 0.0 1.4 The actual concentrations of solution added to each tube are represented by sediment-free controls (#21-#26 in Figure 15). Control selenite concentrations were not identical to theoretical concentrations and fluctuated above or below the predicted values. Figure 16 demonstrates the relationship between theoretical and control selenite concentrations. 103 Figure 16 - Bar graph showing relationship of theoretical to control selenite concentrations. 250 200 1 50 50 H 500 250 62.5 • [ S e j t h e o r e t i c a l ( p p b ) • [ S e ] c o n t r o l ( p p b ) 7 . 1 8 .4 31 .25 15.625 [ N a 2 S e 0 3 ] ( p p b ) For selenate adsorption trials, control Se concentrations were all found to be lower than theoretical concentrations. This was especially significant for the highest selenate concentration; the control selenate concentration (126 ppb) was approximately 60% of the theoretical value (209 ppb). Figure 17 demonstrates the relationship between theoretical and control selenate concentrations. 104 Figure 17 - Bar graph showing relationship of theoretical to control selenate concentrations. 2 5 0 T — — > 5 0 0 2 5 0 1 2 5 6 2 . 5 3 1 . 2 5 1 5 . 6 2 5 [ N a 3 S e O 4 ] ( p p b ) All selenite adsorption treatments (containing sediment) exhibited a marked reduction of selenite in solution (Figure 18). Removal of selenite ranged from 78% -100% in solution; however there was no significant difference in final SelV concentrations between different treatment groups. 105 Figure 18 - Selenite concentrations of controls (no sediment) and treatments (sediment added). • Control [Se] 2225 228.3 114.2 57.1 28.5 14.3 7.1 Theoretical Se Treatment Concentration (ppb) Very different results were found for selenate adsorption trials. Almost all trials had higher final Se concentrations after shaking with sediment (Figure 19). Further analysis of these results can be found in the following discussion section. 106 Figure 19 - Selenate concentrations of controls (no sediment) and treatments (sediment added). 107 5.4. DISCUSSION Theoretical concentrations of selenite and selenate were calculated based on the addition of 0 to 500 pg/L sodium selenite or sodium selenate (refer to Table 12). These concentrations corresponded to theoretical selenite and selenite concentrations found in Tables 13 and 14, respectively. Measured values for the selenite controls were more or less equivalent to the theoretical values, however there was a significant difference between the measured and theoretical selenate values. There is some consistency in the difference between the theoretical and measured selenate controls; for four of the concentrations measured, concentrations were about half of what was expected. This may indicate some dilution error while samples were being prepared for analysis. Nevertheless, these results do not effect adsorption experiment conclusions since measured controls were compared to measured treatments and all samples were treated the same during analysis. Interestingly, the majority of selenate treatments (all except Na 2Se0 4 = 250 ppb) that were measured after the adsorption process, were actually higher in concentration than control treatments. This would indicate possibly some selenate leaching out of sediment; this is further evidence that a more thorough washing procedure may be required for future studies of this type. Figure 18 shows that the concentration of selenite in solution decreases significantly following the adsorption (shaking) process, while Figure 19 did not demonstrate a reduction in soluble selenate. This is consistent with previous research on the transport behavior of selenate. Guo et al. (1999a) found that the mobility of selenate was very high in all soil columns tested (organic amended and unamended soil). These results are in agreement with the literature that states selenate is not sorbed by soil (Neal and Sposito, 1989; Fio et al., 1991) and is readily leached from soil columns (Ahlrichs and Hossner, 1987; Alemi et al., 1988, 1991). This non-sorbing characteristic of selenate indicates that this species is subject to a high leaching potential. As demonstrated in the previous batch microcosm study (Chapter 4), the mobility of selenate is significantly lowered when it is reduced to elemental selenium. Elradashi et al. (1987) worked out theoretically the equilibrium reactions and the associated constants for a wide range of selenium minerals and found that metal selenite/selenate minerals (with one exception MnSeOs) were too soluble and unstable to exist in most soils. However, in alkaline soils ferric oxide-selenites may play an important role in controlling selenite solubility (Neal, 1995), As discussed previously (Guo et al., 1999a), iron oxides such as goethite can adsorb selenite by forming inner and outer sphere complexes (Sposito et al., 1988). Inner-sphere complexes are formed when an aqueous ligand exchanges for a surface hydroxyl group. This is described as specific adsorption. Firstly, the surface is protonated by a proton from the diprotic acid (also the source of the 109 ligand itself) and then the water molecule is exchanged for the ligand, i.e. selenite. The bonding in the inner-sphere complex is either ionic or covalent, and sometimes can be a combination of both. It is this type of bonding that gives stability to selenite adsorption in sediment (Neal et al., 1987). Both selenate and selenite can adsorb to metal hydrous oxides (Zhang and Sparks, 1990). Selenate adsorbs through electrostatic attraction, and selenite through ligand exchange. Both of these processes are very pH dependent. In the case of selenate adsorption to goethite the amount adsorbed decreases to zero at a pH of approximately 7.5, the point of zero charge (PZC) of goethite. Selenite adsorption also drops with increasing pH, however it still adsorbs at pH's up to a pH of 10. We did not measure the pH of our solutions, but we estimate a neutral or slightly alkaline pH. Our results confirm the previous study (Zhang and Sparks, 1990), since no selenate adsorbed while almost all of the selenite did adsorb. 5.5. ADSORPTION CONCLUSIONS For the concentrations used in this study, enough selenite was adsorbed by sediment to significantly reduce aqueous concentrations to within CCME guidelines (5-ppb Se). Thus, Goddard Wetland sediment has sufficient 110 adsorptive capacity for selenite. Selenate, however, did not adsorb at all, and would therefore need to be reduced to selenite to adsorb to sediment. The batch microcosm and adsorption studies have demonstrated that Goddard Wetland sediment has the potential to remove selenium oxyanions from solution: selenite through adsorption, and selenate through bacterial reduction. The next step in this study was to determine to what extent this happens in semi-continuous microcosms, approximately simulating actual conditions in the field. This study is presented in the following Chapter. I l l 6. CONTINUOUS MICROCOSM STUDY 6.1. INTRODUCTION The previous batch microcosm and adsorption studies provided evidence that Goddard Pond sediment is capable of reducing soluble selenite and selenate concentrations. These processes include: a) microbial reduction of selenate to elemental selenium by sulfate-reducing bacteria and selenate-reducing bacteria present in the sediments, and b) adsorption of selenite to sediment. The objective of the continuous microcosm study outlined in this chapter is to assess whether these biogeochemical processes (occurring in batch microcosm and adsorption studies above), would persist in a system representative of Goddard Wetland conditions. These conditions included; an overlying water matrix with 900 mg/L sulphate (as Na2S04), an alkaline pH (between 8.0 and 10.0), and surface airflow. Organic amendments were also added (as Goddard Site #1 vegetation and/or lactic acid with generic plant fertilizer). These amendments were added in order to: a) simulate annual die-back of wetland vegetation, and 112 b) determine if amendments could speed up selenium uptake. These microcosms were run in a semi-continuous fashion to simulate an open system. After 10 mL of sample was removed weekly, 10 mL of 900 mg/L sulfate and 500 ug/L Se solution was added. 6.2. MATERIALS AND METHODS Continuous microcosm protocol was adapted from methods used by Zhang and Moore (1997), Stork, Jury, and Frankenberger (1999), and Zhang and Frankenberger (1999). Our study consisted of nine different treatments, based upon a three-level, two-treatment factorial design, replicated twice (32 • 2 = 18); a total of 18 microcosms were required. This factorial design is outlined in Table 15. 113 Table 15 - Continuous microcosm factorial design where: for lactic acid (LA) and fertilizer (fert), -1 = none 0 = 0.17 mL LA + Idropfert 1 = 0.34 mL LA + 2 drops fert for vegetative matter -1 no vegetation 0 0.5 cm layer vegetation 1 1.0 cm layer vegetation Microcosm ID Lactic Acid + Fertilizer Vegetative Matter 1(a) and (b) -1 -1 2(a) and (b) -1 0 3(a) and (b) -1 1 4(a) and (b) 0 -1 5(a) and (b) 0 0 6(a) and (b) 0 1 7(a) and (b) 1 -1 8(a) and (b) 1 0 9(a) and (b) 1 1 114 6.2.1. EXPERIMENTAL METHODS Site #1 sediment samples utilized in this study were collected and preserved as outlined in Materials and Methods Chapter 3.1.1. Vegetation samples were collected from Goddard Site #1 at the same time as sediment collection. Vegetative sampling involved removing common cattail and beaked sedge from area surrounding Goddard Site #1 water. These samples were sealed in a large Ziploc bag and preserved as soon as possible (within 36 hours of collection) by cooling to and maintaining a temperature of approximately 4°C (ice cold) in a large refrigerator. Continuous microcosms were constructed using 500 mL glass mason jars that had been previously sterilized for food preservation purposes. Lids to mason jars had been fashioned with a spout for out-flowing air and a hole drilled for in-flowing air, sampling, and injecting solution (Figure 20a,b). 115 Figure 20 a, b - Schematic diagram of: a) continuous microcosm, and b) activated carbon column (where CP = cotton plug and AC = activated carbon), a) b) 116 Each microcosm consisted of 0.175 L of Goddard Site #1 sediment, 0.205-L of sulphate solution, an initial selenate spike, and different amendments (as outlined in Table 16). A description of experiment preparation in the laboratory and anaerobic chamber is detailed below. Table 16 - Treatment descriptions for all continuous microcosms. Treatment Sediment so/' Solution Initial Se addition Lactic acid + fertilizer Organic Matter 1a 0.175 L 0.205 L 100 (JL - -1b 0.175 L 0.205 L 100 pL - -2a 0.175 L 0.205 L 100 uL - 0.5-cm layer 2b 0.175 L 0.205 L 100 uL - 0.5-cm layer 3a 0.175 L 0.205 L 100 uL - 1.0-cm layer 3b 0.175 L 0.205 L 100 ML - 1.0-cm layer 4a 0.175 L 0.205 L 100 pL 0.17 mL + 1 drpfert -4b 0.175 L 0.205 L 100 uL 0.17 m L + 1 drpfert -5a 0.175 L 0.205 L 100 ML 0.17 mL + 1 drpfert 0.5-cm layer 5b 0.175 L 0.205 L 100 LIL 0.17 m L + 1 drpfert 0.5-cm layer 6a 0.175 L 0.205 L 100 uL 0.17 mL + 1 drpfert 1.0-cm layer 6b 0.175 L 0.205 L 100 uL 0.17 m L + 1 drpfert 1.0-cm layer 7a 0.175 L 0.205 L 100 uL 0.34 mL + 2 drp fert -7b 0.175 L 0.205 L 100 pL 0.34 mL + 2 drp fert -8a 0.175 L 0.205 L 100 pL 0.34 mL + 2 drp fert 0.5-cm layer 8b 0.175 L 0.205 L 100 ul_ 0.34 mL + 2 drp fert 0.5-cm layer 9a 0.175 L 0.205 L 100 uL 0.34 mL + 2 drp fert 1.0-cm layer 9b 0.175 L 0.205 L 100 LiL 0.34 mL + 2 drp fert 1.0-cm layer 117 In the laboratory, vegetation collected from Goddard Site #1 was prepared by using scissors to cut/homogenize vegetation until a consistent mixture was observed. Four litres (2 X 2L flasks) of 900 mg/L sulphate solution was prepared by adding 2.66 g Na2S04 into each of two flasks each containing 2 L of distilled H 20. This solution was deoxygenated by bubbling N g a s through each flask for i approximately 10 minutes. In an anaerobic chamber, the Goddard sediment sample (collected from Site #1 as discussed in Chapter 3.1.1.) was opened and homogenized (mixed) thoroughly in a large basin using scissors and large plexiglass rods. Scissors were used to cut up any larger debris (i.e. roots and leaves), and very large debris that could not be broken down (i.e. chunks of wood) was removed. While in the anaerobic chamber, 175 mL of sediment sample (measured using 175 mL scoop) was placed into each of 18 microcosms (500 mL glass mason jars). Also in the anaerobic chamber, 200 mL of 900 mg/L sulphate solution (as sodium sulphate) was added into each microcosm jar. After addition of sediment and sulphate solution, all microcosms were sealed and allowed to settle for 72 hours. 118 After 72 hours of settling, 4.95 mL of overlying solution was sampled from each microcosm. Each sample was then filtered (through 0.22 jim filter paper), and preserved with 0.05 mL (50 u.L) H N O 3 in a 15 mL polypropylene Falcon tube (final H N O 3 concentration = 1%). This will be the initial sample (Sample 0 prior to Se-spike) taken from each microcosm for Se analysis. A concentrated selenate solution was prepared to spike each microcosm. 0.024 g of Na 2Se0 4 was dissolved in 10 mL of deoxygenated sulphate solution (prepared above). This is equivalent to 0.01 g of Se(VI) in 10 mL of solution. 100 uL of this Se solution was added to each microcosm. Therefore, assuming that this Se spike is the only source of selenium in each microcosm, the theoretical initial concentration of Se(VI) in each microcosm is 500 ug/L. A generic growth media was prepared using a concentrated generic plant fertilizer and lactic acid. Based on the Thauera selenatis growth medium recipe (outlined in Chapter 4.2.1. by Oremland et al., 1994), this recipe calls for 1.7 mL/L lactic acid. Continuous microcosm treatments were prepared by adding the following amounts to the appropriate microcosms (see Table 11); 0.34 mL lactic acid with two drops plant fertilizer were added to microcosms 7 - 9, and 0.17 mL lactic acid with one drop plant fertilizer was added to microcosms 4 - 6 . The appropriate amount of organic matter (homogenized wetland vegetation) was added to microcosms 2, 3, 5, 6, 8, and 9. By pre-measuring with 119 a 25 mL glass beaker, it was determined that loosely packing the beaker with vegetation (to 25 mL mark) equated to a 1 cm layer of organic matter in microcosm. By loosely packing this beaker to 12.5 mL mark, a 0.5 cm layer of organic matter was obtained in appropriate microcosm. After the addition of organic matter, the pH of each microcosm was measured and all microcosms were sealed and left to settle. One hour after the measurement of pH, a 10 mL sample of overlying water was taken from each microcosm. This sample was filtered, preserved with 0.1 mL nitric acid and refrigerated for future Se analysis (Sample 1 following Se spike). The 10-mL sample that was withdrawn from each microcosm was replaced with 10-mL of top-up solution. Top-up solution was made fresh every week using the following recipe: 0.024 g Na2Se04 was dissolved in 100 mL distilled H 20. 50 uL of this solution was added to 9.95 mL of 900 mg/L sulphate solution (as Na 2S0 4). This 10 mL solution was then replaced into each microcosm. After the addition of top-up solution, the flow of air through each microcosm was opened. Air flow was achieved by connecting the out-flowing air spout to a carbon column; this carbon column was then attached to a vacuum system. Carbon columns were constructed by packing 50 mL polypropylene syringes with activated carbon between two layers of cotton (Figure 20b). 120 The vacuum system was connected to all microcosms, and air flow through each microcosm was adjusted so that any gases volatilized from microcosms were taken up into that microcosm's corresponding carbon column. The vacuum system was set to provide very little suction (less than 0.25 L/minute) from each microcosm. Also, the only time that air flow was turned off was during weekly sampling and topping up of microcosms. All microcosms were kept in the dark, all at the same temperature for the duration of the experiment. Every week (for 12 weeks), 10 mL of overlying water was removed from each microcosm using a volumetric pipette. This sample solution was preserved with HNO3 (1% acid preservation as above) in a 15 mL polypropylene Falcon tube for future Se analysis. After sampling, a fresh 10 mL top-up of selenate-sulphate solution (500 pg/L Se - 900 mg/L SO4 2") was replaced into each microcosm. During the first three weeks of this study, evaporation occurred to differing extents in all microcosms, therefore samples collected over this period were not analyzed. After these three weeks and prior to sampling, each microcosm was topped up to its original volume with deoxygenated, distilled water. Therefore, following this top-up with water each microcosm was sampled for 10 mL of overlying water, and this sample was then replaced by 10 mL of selenate-sulphate top-up solution. 121 After nine weeks, since the Se concentration in the overlying water was very low, all microcosms were re-spiked with 100 jiL of concentrated selenate solution (as mentioned above). Also at this time, lactic acid and plant food were added to the appropriate microcosms as before. Once again, an initial sample (20/12/01) was taken right after the second Se spike. For the following three weeks, each microcosm was topped up every week to its original volume with deoxygenated, distilled water. As before, following this top-up each microcosm was sampled for 10 mL of overlying water, and this sample was then replaced by 10 mL of selenate-sulphate top-up solution. Throughout the duration of this experiment, pH testing was performed for all continuous microcosms once a week for weeks 2, 4, and 6. 122 6.2.2. ANALYTICAL METHODS TOC A Leco CN-2000 Carbon and Nitrogen Analyzer was used to analyze the TOC and Nitrogen in homogenized wetland vegetation. Three replicates samples of organic matter were prepared for: a) 25 mL (1 cm layer)of homogenized wetland vegetation, and b) 12.5 mL (0.5 cm layer) of homogenized wetland vegetation These samples were prepared similarly to sediment samples as discussed in Chapter 3.1.2. (application notes for this instrument can be found in Appendix A). pH Testing A standard mobile pH meter (Mettler Toledo - Type 1140) was used to measure the pH in each microcosm. At the beginning of the experiment, the pH was adjusted to 8 using NaOH or HCI (as necessary). 123 SELENIUM ANALYSIS (HG-AF) Total selenium concentration was determined for solution from each continuous microcosm for each week sampled. This analysis was performed using the Atomic Fluorescence System - Millenium Excalibur PSA 10.055. An in-depth description of this equipment, the analytical methods used to prepare samples, and the determination of total soluble Se using HG-AF can be found in Chapter 4.2.3. 6.3. RESULTS TOC Analysis of carbon and nitrogen in vegetation samples indicate homogeneity among organic matter replicates. This is indicated by the relatively small standard deviations occurring among replicates (Table 17). 124 Table 17 - Sample A = 25-mL (1-cm layer)of vegetation analyzed for TOC and N, Sample B = 12.5-mL (0.5-cm layer)of vegetation analyzed for TOC and N. Sample # Carbon(% dry weight) Nitrogen (% dry weight) A1 43.4 1.5 A2 43.1 1.3 A3 45.3 1.0 A v e A 43.9 (±1.2) 1.3 (±0.2) B1 43.3 1.4 B2 43.0 1.4 B3 43.0 1.4 Avee 43.1 (± 0.2) 1.4 (±0.02) pH MONITORING On all three days for which pH values were monitored, continuous microcosms all fell within the desired pH range of 8 -10 (Table 18). No significant patterns were observed from one pH testing to the next, or among different treatments. 125 Table 18 - pH monitoring of continuous microcosms Microcosm # Week Two Week Four Week Six 1a 8.0 (8.2) 8.2 8.1 1b 8.2 8.2 (8.0) 8.0 2a 8.2 8.3 8.0 2b 8.2 8.3 8.1 3a 8.3 8.4 8.3 3b 8.3 8.5 8.8 (8.8) 4a 8.6 8.6 (8.6) 8.0 4b 8.6 8.6 8.1 5a 8.4 8.6 8.2 5b 8.3 8.5 8.6 6a 8.4 8.7 9.3 6b 8.3 8.9 9.7 (9.6) 7a 8.6 8.8 8.4 7b 8.7 8.7 8.5 8a 8.3 (8.2) 8.9 8.5 8b 8.4 (8.4) 8.6 8.4 9a 8.4 8.4 8.4 9b 8.3 8.4 8.4 Values in brackets indicate duplicates 126 HG-AF SELENIUM ANALYSIS As seen in Figure 21, Se concentrations in all microcosms decreased markedly following spiking. Se spikes were added to each microcosm on week zero and week nine. The purpose of the Se spike was to bring the concentration in each microcosm up to 500 ug/L, after which Se concentrations over time could be monitored to see if Se was being removed from the water. The second spike was added in week nine because the Se concentrations were low and we wished to see if the observed uptake of selenate could be replicated. A control microcosm was included in this study; this microcosm consisted only of sulphate solution (no sediment), and received the same Se spike as all other microcosms. Se concentration in the control remained relatively constant from sample week ten (464 ppb Se) to sample week 12 (439.2 ppb Se) (Table 20). Samples were collected right after spiking each microcosm with Se. Samples from week zero ranged from approximately 154 ppb to 697 ppb total Se, while samples from week nine ranged from 178 ppb to 573 ppb. Analytical accuracy was confirmed by running replicates on different samples on different days, all yielding consistent results. 127 Table 19 - Individual continuous microcosm treatment Se concentrations (ppb). Treatment WeekO Week 4 Week 5 Week 9 Week 10 Week 11 Week 12 1a ns 27.00 5.43 320.05 51.81 48.44 29.34 1b 541.94 8.24 10.00 231.38 50.62 47.65 16.17 2a 515.58 24.64 5.08 532.68 29.63 34.98 17.36 2b 468.20 9.69 1.69 284.97 33.20 32.11 17.36 3a 697.04 63.44 7.31 536.36 11.42 18.74 13.30 3b 447.25 12.71 5.32 505.20 17.56 30.23 15.92 4a 191.10 70.07 5.43 216.43 62.21 54.38 18.45 4b 195.49 7.47 5.90 293.83 48.59 48.59 25.08 5a 391.91 6.76 15.04 365.01 41.86 36.22 12.11 5b 369.95 14.00 7.54 317.63 27.75 30.23 30.62 6a 554.42 28.43 4.61 281.98 25.77 24.88 14.59 6b 443.02 112.57 3.68 573.05 34.04 28.74 13.10 7a 154.03 64.09 7.89 178.02 101.81 39.78 89.43 7b 420.15 25.57 12.58 287.62 69.14 29.93 57.35 8a 295.05 27.64 8.60 ns 33.50 21.12 22.11 8b 316.54 90.54 8.01 310.96 18.65 29.14 12.61 9a 355.82 14.60 9.77 ns 21.62 26.61 17.80 9b 321.12 68.76 7.78 367.43 43.20 24.88 57.35 ns = no sample analyzed Table 20 - Average continuous microcosm treatment Se concentrations (ppb). Treatment WeekO Week 4 Week 5 Week 9 Week 10 Week 11 Week 12 1 541.9 17.6(13.3) 7.7 (3.2) 275.7 (62.7) 51.2 (0.8) 48.1 (0.6) 22.8 (9.3) 2 491.9 (33.5) 17.2 (10.6) 3.4 (2.4) 408.8(175.2) 31.4 (2.5) 33.5 (2.0) 17.4 (0.0) 3 572.2(176.6) 38.1 (35.1) 6.3(1.4) 520.8 (22.0) 14.5 (4.3) 24.5 (8.1) 14.6 (1.9) 4 193.3 (3.11) 38.8 (44.3) 5.7 (0.3) 255.1 (54.7) 55.4 (9.6) 51.5 (4.1) 21.8 (4.7) 5 380.9(15.5) 10.4 (5.12) 11.3(5.3) 341.3 (33.5) 34.8 (10.0) 33.2 (4.2) 21.4(13.1) 6 498.7 (78.8) 70.5 (59.50) 4.2 (0.7) 427.5 (205.8) 29.9 (5.9) 26.8 (2.7) 13.8 (1.1) 7 287.1 (188.2) 44.8 (27.2) 10.2 (3.3) 232.8 (77.5) 85.5 (23.1) 34.9 (7.0) 73.4 (22.7) 8 305.8(15.2) 59.1 (44.5) 8.3 (0.4) 311.0 26.1 (10.5) 25.1 (5.7) 17.4 (6.7) 9 338.5 (24.53) 41.7(38.3) 8.8 (1.4) 367.4 32.4 (15.3) 25.8(1.23) 37.6 (28.0) Control - - - 464.2 494.8 439.2 Shaded samples indicate samples taken after Se spike on same day Values in brackets indicate standard deviations 128 Figure 21 a, b - Average continuous microcosm treatment for total selenium in overlying water samples. a) Week 0 (spike) - Week 5 8 0 0 6 0 0 a j3 § 4 0 0 c o O m 2 m > < 2 0 0 ttL • W e e k 0 • W e e k 4 • W e e k 5 b) Week 9 (spike) - Week 12 T r e a t m e n t # 800 600 Xi Ch cz o C § 400 c o O 5 200 4 • w e e k 9 • w e e k 1 0 • w e e k 1 1 • w e e k 1 2 T r e a t m e n t # 129 6.4. DISCUSSION The carbon and nitrogen analysis performed with samples of wetland vegetation, indicate that these organic amendments were well homogenized and provided a consistent source of carbon and nitrogen to those treatment microcosms (Table 17). Also, no significant trends were observed for pH testing of different treatments on different days. On the three days for which pH values were monitored, continuous microcosms all fell within the desired pH range of 8 -10 (Table 18). pH was monitored periodically to ensure that conditions in each microcosm were similar to the slightly alkaline conditions (pH 8 - 1 0 ) observed in Goddard Wetland. The most important analysis to be performed in this study was selenium determination in overlying water samples. These samples were taken to determine whether selenium (as selenate) was removed from solution, and whether different amendments, and quantities thereof, would impact the rate of Se removal from overlying water. Although initial samples (those taken on days of Se spiking), had higher Se concentrations, the results appeared to be variable both among treatments and between treatment replicates. For example, week zero samples from 130 treatments 7a and 7b taken one hour after spiking, were found to have very different Se concentrations (154 ppb and 420 ppb, respectively). A concentration of 500-ppb was expected, or greater given that the original sediment samples were wet and are composed of Se-containing coal. Some Se concentrations are within the expected range, however quite a few are less (up to half) that of the expected 500 ppb. Therefore the treatments for which replicates were not obtained, are possibly inaccurate and/or not representative of actual treatment results. For future studies of this type, it is recommended that several replicates (i.e. n = 5) be run for each treatment type. Many possible reasons exist for the variability in initial Se concentrations: 1) This variability may have been due to other sources of Se contamination. It was however concluded that this was not the case, since many microcosms contained less than the expected 500-ppb Se. Few microcosms contained concentrations that were greater than 500-ppb. 2) Adsorption of Se to the microcosm glass may have occurred. This may have not been the case since the control had a constant Se concentration very close to what was expected. Furthermore, the previous adsorption study indicates that selenate does not adsorb to the sediment. 3) Volatilization / biogeochemical reduction of Se by microorganisms in sediment and/or organic matter, may have contributed to the variability of initial Se 131 concentrations. Given the results of the batch microcosm experiments, it is possible that microbial reduction of selenate occurred. However, it is unlikely that this would occur in the hour between adding the Se spike and taking the first sample. 4) Errors during Se analysis may have occurred, however the analytical procedure used for Se proved to be very reliable. Samples were re-analyzed on different days and consistent results were obtained. Zeros and standards always gave the expected results. 5) Inaccuracies in sampling procedure may have been the most probable cause of variability in initial Se concentrations. Microcosms were not thoroughly mixed, therefore the 100-ul Se spike solution may not have mixed very well with the rest of the solution in the jars. Perhaps in future experiments, Se should be added to the 175-mL sulfate solution and mixed well before adding the solution to the microcosms. This could only be done at the beginning of the experiment during set-up. Alternatively, gentle mixing of overlying solution after the addition of the Se spike or top-up solution, might contribute to a more consistent application of selenium to microcosms. 132 Given the success of the batch microcosm experiments, biogeochemical reduction and volatilization of selenate were the processes we hoped to observe with this study. Reduction was to be observed qualitatively (the presence of red precipitate) and/or quantitatively (by HG-AF Se analysis). Volatilization was to be assessed quantitatively with the digestion and analysis of carbon columns for adsorbed volatile Se. This latter portion of the study was not possible due to time and budgetary constraints. A control microcosm (prepared without sediment) provided consistent results near 500 uxj/L for the last three weeks of this study. In contrast to the control, all treatment microcosms showed a marked decrease in selenium concentration in the weeks after the two Se spikes. Figures 21 a, b illustrate the reduction of Se in overlying solution in the weeks following selenium spikes. The consistent control (sediment-free) Se concentrations indicate that selenate was removed from solution by sediment and/or amendments. Since adsorption experiments demonstrated that selenate does not adsorb to (sterilized) sediments, we postulate that microbial processes in the sediment were responsible for selenate removal, however we were surprised by the rapidity with which this occurred (within one week approximately 90% of the Se added was removed from solution). This postulation needs to be confirmed by a follow-up analysis of sediment and carbon columns for each microcosm. For example, SEM-EDX analysis of sediment could help to identify the predominant 133 species of selenium present in sediment; or the digestion and analysis of activated carbon could tell us just how much selenium is volatilized from each continuous microcosm. Due to time and economical constraints, these analyses were not performed for this continuous microcosm study. As illustrated in Figure 21 a, b, large standard deviations made a comparison of different treatment difficult to perform. However, a trend indicating that increased organic amendment (vegetation) resulted in increased Se concentration was observed after spiking. This hypothesis is based upon the following: 1) Treatments 4 and 7 seemed to have the lowest Se concentrations of all treatments for weeks 0 and 9. Note that neither of these treatments were amended with vegetative matter; this may indicate that vegetation was an actual source of Se in continuous microcosms. 2) Treatments 4, 5, 6, and treatments 7, 8, 9 seem to show a gradual increase in Se concentration corresponding to stepwise increases in vegetative amendment. This further substantiates the hypothesis that vegetative amendments were an introduced source of Se. Analyses of variance were performed comparing treatments at the end of the experiment; all P-values exceeded the significance level of 0.05 for all weeks, 134 and could therefore not be considered significantly different. For future study, running several replicate microcosms, and collecting a larger number of samples might help to eliminate outliers and provide evidence of trends: 6.5. CONTINUOUS MICROCOSM CONCLUSIONS The objective of the continuous microcosm study was to assess whether the biogeochemical processes (occurring in earlier batch microcosm and adsorption studies above), would persist in an open (continuous) system representative of Goddard Wetland conditions. The results clearly demonstrated that Se concentrations in solution decrease markedly after spiking. Some problems with the methodology produced different (or uncertain) initial Se concentrations. However, Se concentrations clearly decreased in all cases. Also, no statistically significant trends were observed among continuous microcosm treatments. However, taking the results of this work into consideration for future continuous microcosm studies may contribute to more functional data being obtained. In conclusion, from information gathered during the adsorption and batch microcosm experiments, we postulate that microbial processes in the sediment were responsible for Se removal. 7. CONCLUSIONS AND RECOMMENDATIONS 135 7.1 CONCLUSIONS Surface coal mining at the Elkview Coal Mine in southeastern British Columbia is a significant source of waterborne selenium entering lentic waterways (i.e. wetlands) in the Elk River Valley. Selenium research in these slow-moving waterways is important because these types of ecosystems are some of the most important feeding and breeding habitats for fish and wildlife. The Goddard wetland is indicative of a slow-moving waterway receiving drainage water directly from the Goddard settling pond on Elkview property. Based on a site survey and characterization, sediment from a small shallow pond (Site #1) closest to the settling pond decant was chosen to be used in Batch Microcosm, Adsorption, and Continuous Microcosm studies. The objective of the batch microcosm study was to test for the presence of selenium reducing bacteria. Qualitative and quantitative monitoring of batch microcosms indicated the potential reduction of selenate out of solution in treatment microcosms containing Goddard sediment. This study indicated potential for the use of growth media application in reducing soluble selenium in the natural environment and/or in remediation technologies. However, this study 136 did not provide any quantitative data on the role of adsorption of selenium oxyanions to sediment. Subsequent adsorption studies with selenate and selenite indicated that selenite was capable of adsorbing to particulates in treatments containing sediment or inoculum. Selenate showed significantly less potential in adsorbing to sediment. Therefore, in retrospect to batch microcosm studies the reduction of selenate (the most mobile form of Se in soils) to selenite, elemental Se, or volatile forms of Se may have contributed to reduced concentrations of Se in solution. The objective of the continuous microcosm study was to assess whether the biogeochemical processes (occurring in previous batch microcosm and adsorption studies), would persist in an open (continuous) system representative of Goddard Wetland conditions. Although all continuous microcosms showed a decrease in soluble Se concentrations in weeks following a selenium spike, variable results indicate that (in addition to reduction and adsorption) several mechanisms may come into play with regards to Se cycling in an open system. 137 7.2. RECOMMENDATIONS The studies performed do not elucidate how different wetland processes compete for soluble selenium in an environment such as the Goddard Wetland. However, all three studies provide valuable information and potential for the development of effective methods for accelerating and optimizing in-situ bioremediation processes (such as reduction or adsorption of Se). The understanding of selenium cycling in aquatic systems is progressing with the application of computer simulation methods. The dynamics of contaminant flows and concentrations are now being provided on many scales, from the small, fast biochemical and microbial processes to the global flows of the biosphere (Odum, 2000). The following outlines two different simulation modeling techniques that could be used for future study of selenium cycling in wetlands. Jorgenson (1979) summarized data from the literature for modeling metal cycling. This type of model could be applied to the study of Se cycling in wetlands. Parameters involved in this model included: uptake by organisms, exchange of toxicants with sediments, exchanges with suspended particles, and exchange with organic substances. Jorgenson (1979, 1995) then used graphs of generation time and decreasing metabolic rates with increasing size of organisms to calibrate rates of biological uptake and release of metals. 138 Thoman (1984) listed the features needed in environmental models of hazardous substances. He also provided parameters for several mechanisms that could potentially occur with regards to Se cycling in wetlands. This system model was a combination of the separate equations for mechanisms listed below: 1. sorption - desorption mechanisms between water and sediment or particles; 2. losses of toxicant through biodegradation, volatilization, chemical reactions, and photolysis; 3. advection transport or dispersion of toxicant; 4. settling mechanisms; 5. external inputs; 6. sorption by organisms; 7. feeding intake by organisms; 8. assimilation into growth of organisms; 9. prey-predator transfers; 10. depuration or excretion by organisms; For future studies, utilizing simulation modeling techniques in concurrence with batch microcosm, adsorption, and continuous microcosm studies could provide valuable information. Results from such studies could ultimately be applied to ex-situ waste treatment alternatives, such as constructed wetlands or biostimulation of natural wetlands. 1 3 9 APPENDIX A - FOLLOWING PAGES CONTAIN LECO CN 2000 ORGANIC APPLICATION NOTES FOR TOTAL / ORGANIC CARBON AND NITROGEN ANALYSIS IN SOILS. m— CN-2000 Total /Organic Carbon and Nitrogen i n Soils Accessaries 528-203 Combustion Boat, 502-343 Nickel Liner, HCI, distilled H.O, hot plate, and 769-608/769-610 Halogen Trap Materials for (non-carbonate carbon procedure), 502-029 Synthetic Carbon Standard Calibration Standard 502-309 Soil, or other suitable standard Sample Weight 0.25 to 0.50 c Analysis Time ~5 minutes Furnace Temperature 135( Oxygen Profile Burn Cycle Lance Flow 1 OFF 2 ON 3 ON 4 ON 5 ON Note: If the lance and purge flow rotometers are set correctly, the time between answer printouts should be five minutes when using the oxygen profile listed above. Procedure for Total Carbon and Nitrogen 1. Prepare instrument as outlined in the operator instruction manual (perform maintenance and leak checks). 2. Analyze blanks until instrument is stable, then analyze three to five boats containing -1.5 g 501-427 COM-AID and three to five empty 528-203 Combustion Boats. Enter 0.2500 g as the weight. Set a blank using results from the empty boats. 3. Weigh -0.25 g of standard material into a 528-203 Combustion Boat and analyze three to five times, then drift correct using these values (refer to operator's instruction manual for details). 4. Weigh -0.25 g of sample into a 528-203 Combustion Boat and analyze. 5. Analyze a standard at the end of the set to verify calibration. OFF END Procedure for Non-Carbonate Carbon A weighed sample is treated then dilute acid to remove carbonate carbon. The treated sample is analyzed to determine the non-carbonate carbon. The difference between the total carbon results and the acid treated results is referred to as the carbonate carbon content. The non-carbonate carbon results (acid treated) are usually referred to as "Total Organic Carbon" (TOC). 606-327 REAGENT TUBS 806-M7 . REAGENT TUBE 1. Prepare instrument as outlined in the operator instruction manual (perform maintenance and leak checks). Install 769-608 and 769-610 Halogen Trap Materials into secondary reagent tube following the diagram on the right. continued on back 501-171 ANHYDRONE 301-081 GLASS WOOL 13mm (0.5 inches) A N T I M O N Y M E T A L 20mm (0.75 inches) 501-081 GLASS WOOL 13mm (0.5inchee) 501-081 GLASS WOOL 13mm ( 0 5 inches) 710-810 F. Cl ABSORBENT 84mm 125 inches) GLASS WOOL 13mm (as Inches) 789-508 ANTIMONY METAL 20mm (0.75 inohee) • 501-0*1 GLASS WOOL 13mm (0.5 inches) 141 2. Weigh 0.25 to 0.5 g of samples into 502-343 Nickel Liners and record the weights (handle liners with tongs). Also weigh -0.25 g of 502-029 Synthetic Carbon. The synthetic carbon standard will be used to check the blank for this acid treatment procedure. a. Using an eyedropper, add 1:1 (HCI: H,0 by volume) solution to samples and synthetic carbon standard until completely wetted. Try to use approximately the same amount of acid on each blank and sample. b. Place on hot plate (low setting) until dry. Remove and cool. Wet the samples again with diluted acid solution. If no reaction occurs, no further acid treatment is required. Dry on hot plate and cool. WARNING: Use extreme caution when handling acid. Follow all appropriate safety precautions, wear adequate protective clothing, and handle only under a fume hood. 3. Place 502-343 Nickel Liner into a 528-203 Combustion Boat and follow procedure below. NOTE: Handle boats and liners with tongs. This is especially important if carbon is <0.5%. 4. Analyze blanks until instrument is stable, then analyze three to five combustion boats containing nickel liners, and three to five empty combustion boats. Enter 0.2500 g as the weight. Set a blank using results from the empty boats. 5. Weigh -0.25 g of standard material into a 528-203 Combustion Boat and analyze three to five times, then drift correct using these values (refer to operator's instruction manual for details). 6. Enter weight of samples from step 2 and analyze. 7. Analyze a standard at the end of the set to verify calibration. Typical Results for Total Carbon/Nitrogen Sample Weight (g) % Carbon % Nitrogen Soil #1 0.2780 13.01 0.037 0.3188 12.90 0.037 0.2940 13.12 0.036 A v e r a g e = 13.01 0.037 S t d . Dev. = 0.09 0.001 Soil #2 0.3978 3.92 0.054 0.3814 3.98 0.055 0.4420 3.93 0.053 A v e r a g e = 3.94 0.054 S t d . Dev. = 0.03 0.001 Soil #3 0.3800 325 0.039 0.4174 3.26 0.039 0.3849 3.33 0.045 A v e r a g e = 3.28 0.040 S t d . Dev. = 0.04 0.002 Typical Results for Non-Carbonate Carbon Sample Weight (g) % Carbon Soil #1 02494 6.10 0.2541 6.13 A v e r a g e = 6.12 S t d . Dev. = 0.02 Soil #2 0.3987 0.80 0.4030 0.78 0.4503 0.78 A v e r a g e = 0.79 S t d . Dev. = 0.01 Soil #3 0.4220 0.50 0.4141 0.51 0.3920 0.51 A v e r a g e = 0.51 S t d . Dev. = 0.01 LECO Corporation • 3000 Lakeview Ave. • St. Joseph, Ml 49085-2396 Phone: 800-292-6141 • Fax: 616-982-8977 • tnfo@leco.com • www.leco.com LECO is a registered trademark of LECO Corporation Form No. 203-821-165 DCP19-REV1 <S 2000 LECO Corporation 142 APPENDIX B - STATISTICAL RESULTS FOR BATCH MICROCOSM S1M3. S1M3 Selenium Analysis - Anova: T1 Vs T3 SUMMARY Groups T1 T3 Count 3 4 Sum 45.4 313.8 Average 15.1 78.5 Variance 0.4 423.8 ANOVA Source of Variation Between Groups Within Groups SS 6.88E+03 1.27E+03 df 1 5 MS 6.88E+03 2.54E+02 F 27.0 P-value 0.003 Fcrit 6.6 Total 8149 6 P-value<0.05 therefore accept hypothesis that T1 and T3 are significantly different S1M3 Selenium Analysis - Anova: T1 Vs T5 SUMMARY Groups T1 T5 Count 3 4 Sum 45.4 2476.1 Average 15.1 619.0 Variance 0.4 2776.5 ANOVA Source of Variation Between Groups Within Groups SS 6.25E+05 8.33E+03 df 1 5 MS 6.25E+05 1.67E+03 F 375.2 P-value 6.764E-06 Fcrit 6.6 Total 633505 6 P-value<0.05 therefore accept hypothesis that T1 and T5 are significantly different 143 S1M3 Selenium Analysis - Anova: T1 Vs T7 SUMMARY Groups T1 T7 Count 3 3 Sum 45.4 64.4 Average 15.1 21.5 Variance 0.4 58.2 ANOVA Source of Variation Between Groups Within Groups SS 6.01 E+01 1.17E+02 df 1 4 MS 6.01 E+01 2.93E+01 F 2.0 P-value 0.226 Fcrit 7.7 Total 177 5 P-value>0.05 therefore reject hypothesis that T1 and T7 are significantly different S1M3 Selenium Analysis - Anova: T2 Vs T4 SUMMARY Groups T2 T4 Count 4 3 Sum 1.1E+06 6.6E+04 Average 2.8E+05 2.2E+04 Variance 4.5E+09 9.9E+07 ANOVA Source of Variation Between Groups Within Groups SS 1.18E+11 1.38E+10 df 1 5 MS 1.18E+11 2.76E+09 F 42.9 P-value 0.001 Fcrit 6.6 Total 1.E+11 6 P-value<0.05 therefore accept hypothesis that T2 and T4 are significantly different 144 S1M3 Selenium Analysis - Anova: T2 Vs T6 SUMMARY Groups T2 T6 Count 4 4 Sum 1.1E+06 1.5E+06 Average 2.8E+05 3.6E+05 Variance 4.5E+09 4.3E+09 ANOVA Source of Variation SS Between Groups 1.27E+10 Within Groups 2.66E+10 df 1 6 MS 1.27E+10 4.43E+09 F 2.9 P-value 0.142 F crit 6.0 Total 4.E+10 7 P-value>0.05 therefore reject hypothesis that T2 and T6 are significantly different S1M3 Selenium Analysis - Anova: T4 Vs T6 SUMMARY Groups T4 T6 Count 3 4 Sum 6.6E+04 1.5E+06 Average 2.2E+04 3.6E+05 Variance 9.9E+07 4.3E+09 ANOVA Source of Variation Between Groups Within Groups SS 2.01 E+11 1.32E+10 df 1 5 MS 2.01 E+11 2.64E+09 F 76.3 P-value F crit 3.261 E-04 6.6 Total 2.E+11 6 P-value<0.05 therefore accept hypothesis that T4 and T6 are significantly different 145 S1M3 Selenium Analysis - Anova: T4 Vs T8 SUMMARY Groups T4 T8 Count 3 3 Sum 6.6E+04 7.4E+03 Average 2.2E+04 2.5E+03 Variance 9.9E+07 5.6E+04 ANOVA Source of Variation Between Groups Within Groups SS 5.71 E+08 1.98E+08 df 1 4 MS 5.71 E+08 4.95E+07 F 11.5 P-value F crit 0.027 7.7 Total 8.E+08 5 P-value<0.05 therefore accept hypothesis that T4 and T8 are significantly different 146 APPENDIX C - QUALITATIVE RESULTS OF PRELIMINARY BATCH MICROCOSM TRIALS S1M1 AND S1M2. TREATMENT 1 - SEDIMENT AND SRB MEDIUM In S1M1, no changes were observed in any of the treatment replicates for 13 days, after which the walls of all serum vials were slightly stained black. For the duration of this trial after day 13, the treatment 1 sediment-media slurry became darker in colour, and an evolution of gas was observed as bubbles in solution and pressure build-up in vials. No formation of elemental selenium (formation of red precipitate) was observed throughout experiment Identical observations were found for S1M2. Reduction of S0 42 " was evident after 12-13 days as a darker colour and evolution of gas. TREATMENT 2 - SEDIMENT AND T. SEL MEDIUM In S1M1, potential reduction of selenate was first observed as a light pink blush in solution eight days after incubation. After this time, solution became a dark red, and a red precipitate was observed forming on the interface between sediment and growth media. This precipitate then began to adhere to walls of serum vials. 147 S1M2 Treatment 2 was different from the two other Batch microcosm trials in that T.sel medium was prepared with 500 ux)/L selenate rather than 3.8 mg/L selenate. No apparent growth of selenate -reducing bacteria, precipitation of selenium, or evolution of gas was observed throughout monitoring of this experiment. TREATMENT 3-AUTOCLAVED SEDIMENT AND SRB MEDIUM In trial S1M1, growth of SRB was evident as dark stains on the walls of serum vials approximately 6 days after incubation. By the end of the monitoring period, slurry had become dark black in colour with SRB growth persisting on the walls of vials. Although some bubbles were observed in solution, evolution of gas was not observed as pressure build-up in vials. Reduction of selenate to elemental Se was not observed in any of the vials. Some growth of SRB was evident in S1M2 serum vials, as solution became darker in colour after eight days of incubation. The end result of this treatment was identical to S1M1 treatment 3 results; some growth of SRB with bubbles occurring in solution, however there was no formation of Se precipitate. 148 TREATMENT 4 - AUTOCLAVED SEDIMENT AND T. SEL MEDIUM Seven days after incubation of S1M1 Treatment 4, a light pink blush was evident on surface of sediment in one of the four replicates. At the end of the experiment, three of the four replicates had exhibited a colour change; two vials maintained a very light pink blush, while a red precipitate formed on the sediment surface of another vial. S1M2 Treatment 4 showed no change in colour, no formation of precipitate, and no evolution of gas. TREATMENT 5 - SRB MEDIUM For all batch microcosm trials, SRB medium remained clear throughout the duration of each experiment. ( TREATMENT 6 - T. SEL MEDIUM For almost all batch microcosm trials, T. selenatis medium remained clear throughout the duration of each experiment. Only one vial (S1M1 Treatment 6a) exhibited slight pink haze with minimal formation of red precipitate. 149 TREATMENT 7 - SRB INOCULUM AND SRB MEDIUM In S1M1, vials became slightly cloudier (darker) 16 days after start of experiment, however no prominent colour changes or evolution of gas was observed throughout this trial. 13 days after start of experiment, all treatment 7 vials in S1M2 became very black and evolution of gas was observed (as bubbles in solution and pressure build-up in vials). TREATMENT 8 - SeRB INOCULUM AND T. SEL MEDIUM In all S1M1 replicates of this treatment, a bright red precipitate formed within solution four to five days after incubation. Between five and eight days after incubation, the red precipitate became brighter (more intense) in colour. After day 8 the red colour became darker and gradually faded to a cloudy reddish-pink (between days 17 and 19). No colour changes or evolution of gas was observed in any of the replicates throughout the duration of S1M2 treatment 8. 150 Figure C1 - Digital photograph monitoring of batch microcosm trial S1M1. 25/05/2001 04/06/2001 10/06/2001 (III mi mi Treatment 1 llll mi un Treatment 2 llll mi un Treatment 3 llll llll llll Treatment 4 151 Figured continued... 25/05/2001 04/06/2001 10/06/2001 \ Treatment 8 152 APPENDIX D - SULPHATE ANALYSIS OF PRELIMINARY BATCH MICROCOSM S1M1 For batch microcosm trial S1M1, sulphate concentrations were analyzed for the concentrated stock salt solution, treatment 5 (SRB medium only), treatment 1 (sediment +SRB medium), and treatment 3 (autoclaved sediment + SRB medium). SRB medium had a concentration of 3224-ppm S0 42 ' , while the stock salt solution had a concentration of more than 12 times that concentration (39 984 ppm S0 42"). Treatment 1 S0 42 " concentrations (N=3) ranged from 788 ppm to 871 ppm with an average concentration of 828 + 42 ppm. Treatment 3 S0 42 " concentrations (N=4) ranged from 866 to 954 ppm with an average concentration of 922 ppm ± 34 ppm. Figure D1 - Summary of sulphate analysis for batch microcosm S1M1. 4000 3224 ppm 0 Range = 786 - 870 ppm Average = 828 ppm Range = 888 - 955 ppm Average = 922 ppm media media + sediment media + autoclaved sediment 153 REFERENCES Adams, D.G., and T.M. Pickett, 1998. Microbial and Cell-free Selenium Bioreduction in Mining Waters. In Environmental Chemistry of Selenium. Frankenberger, W.T., and R. Engberg (eds.). Marcel Dekker, Inc. New York, 479-500. Ahlrichs, J.S., and L.R. Hossner, 1988. Selenate and selenite mobility in overburden by saturated flow. Journal of Environmental Quality. 16:95-98. Alemi, M.H., D.A. Goldhamer, and D.R. Nielson, 1988. Selenate transport in steady-state water-saturated soil columns. Journal of Environmental Quality. 17: 608-613. Alemi, M.H., D.A. Goldhamer, and D.R. Nielson, 1988. Modeling selenium transport in steady-state, unsaturated soil columns. Journal of Environmental Quality. 20: 89-95. Allaway, W.H., 1968. Control of the environmental levels of selenium. In: Trace Substances and Environmental Health. Volume 12. Hemphill, D.D. (Ed.), University of Missouri, Columbia, Mo., 181. Allison, L.E., 1965. Organic Carbon, in Methods of Soil Analysis. Black, C.A., Evans, D.D., White, J.L., Clark, F.E. (Eds.). American Society of Agronomy, Inc. Madison, Wl . Barnes, J.M., J.K. Polman, and J.H. McCune, 1992. Selenate reduction by Pseudomonas stutzeri. Abstract of the Annual Meeting of the American Society of Microbiology, 384-387. Bennett, W.N., 1984. Examination of selenium kinetics and transfer within a defined aquatic food chain. MS Thesis. Dept. of Zoology, University of Wisconsin-Milwaukee, 55-61. Bradley, P.M., J.E. Landmeyer, and F.H. Chapelle, 1999. Aerobic Mineralization of MTBE and tert-Butyl Alcohol by Stream-Bed Sediment Microorganisms. Environmental Science and Technology. 33: 1877-1879. Bradley, P.M. and F.H. Chapelle, 1997. Kinetics of DCE and VC Mineralization under Methanogenic and Fe(lll)-Reducing Conditions. Environmental Science and Technology. 31 : 2692-2696. Bronzetti, G., E. Morichetti, C. Salvadori, C. della Croce, 1993. Studies of antimutagenic factors in food. Food and cancer prevention: chemical and biological aspects. Royal Society of Chemistry (RSC),Cambridge, UK. 23:180-184. Brown, A T . , and A. Shrift, 1980. Assimilation of selenate and selenite by Salmonella typhimurium. Canadian Journal of Microbiology. 28:307-310. Bryant, R.D., and E.J. Laishley, 1988. Evidence for two transporters of sulfur and selenium oxyganions in Clostridium pasteurianum. Canadian Journal of Microbiology. 34:700-703. Calderone, S.J., W.T. Frankenberger, D.R. Parker, and U. Karlson, 1990. Influence of temperature and organic amendments on the mobilization of selenium in sediments. Soil Biology and Biochemistry. 22:615-620. Casey, R., and Siwik, P., 2000. Concentrations of selenium in surface water, sediment, and fish from the McLeod, Pembina and Smoky Rivers. Interim Report. Natural Resources Service, 1-25. Combs, G.F., and S.B. Combs, 1986. The Role of Selenium in Nutrition. Academic Press, Orlando, 532 p. 154 Corns, W.T., P.B. Stockwell, L.C. Ebdon, and S.J. Hill, 1993. Development of an atomic fluorescence spectrometer for the hydride forming elements. Analytical Atomic Spectrometry. 8: 71-72. Cutter, G.A., and K.W. Bruland, 1984. The marine boigeochemistry of Se: A re-evaluation. Limnological Oceonagraphy. 29: 1179-1192. Demoll-Decker, H., and J. Macy, 1993. The periplasmic nitrite reductase of Thauera selenatis may catalyze the reduction of selenite to elemental selenium. Archives of Microbiology. 160:241-247. DMGL - Digital Mapping Group Limited, 1989. Map prepared for BC Ministry of Environment and Parks. Terrain Resource Information Management, Mao 82G.076. Dobbs, M. G., D. S. Cherry, et al., 1996. Toxicity and bioaccumulation of selenium to a three-trophic level food chain. Environmental Toxicology and Chemistry. 15(3): 340-347. Doran, J.W., 1982. Microorganisms and the biological cycling of selenium, in Advances in Microbial Biology, Vol.6 (K.C. Marshall, Ed.), Plenum, New York, pp.1-31. Dreher, G.B., and R.B. Finkelman, 1992. Selenium mobilization in a surface coal mine. Powder River Basin, Wyoming, USA. Environmental Geological Water Science. 19: 155-167. Elrashidi, M.A., D.C. Adriano, S.M. Workman, W.L. Lindsay, 1987. Chemical equilibria of selenium in soils: a theoretical development. Soil Science. 144: 141-151. Fio, J.L., R. Fujii, and S.J. Deverel, 1991. Selenium mobility and distribution in irrigated and nonirrigated alluvial soils. Journal of the Soil Science Society of America. 55, 1313-1320. Frankenberger, W.T., and U. Karlson, 1989. Microbial Volatilization Of Selenium At Kesterson Reservoir. Journal of the Soil Science Society of America. 53: 1435-1438. Frankenberger, W.T., and U. Karlson, 1994. Microbial Volatilization of Selenium from Soils and Sediments. In Selenium in the Environment. Ed. W.T. Frankenberger. Marcel Dekker, New York, 369-388. Frost, R.R. and Griffin, R.A., 1977. Effect of pH on adsorption of arsenic and selenium from landfill leachate by clay minerals. Journal of the Soil Science Society of America. 4 1 : 53-56. Greenberg, A., L. Clesceri, and A.D. Eaton (eds), 1995. Standard methods for the examination of water and wastewater. American Public Health Association, Washington, DC. Guo, L, W.T. Frankenberger, and W.A. Jury, 1999a. Evaluation of simultaneous reduction and transport of selenium in saturated soil columns. Water Resources Research, 35. 3: 663-669. Guo, L., W.T. Frankenberger, and W.A. Jury, 1999b. Adsorption and Degradation of Dimethyl Selenide in Soil. Environmental Science and Technology, 33. 17: 2934-2938. Hansen, D., Duda, P., Zayed, A., Terry, N., 1998. Selenium removal by constructed wetlands: role of biological volatilization. Environmental Science and Technology. 32: 591-597. Haygarth, P.M., 1994. Global Importance and Global Cycling of Selenium. In Selenium in the Environment. Ed. W.T. Frankenberger. 1994. Marcel Dekker, New York, 1-28. 155 Huang, P.M., and R. Fujii, 1996. Selenium and Arsenic in Methods of Soil Analysis. Part 3 -Chemical Methods. Soil Science Society of America Book Series, 793-1011. Jorgenson, S.E., 1979. Modeling the distribution and effect of heavy metals in aquatic ecosystems. Ecological Modelling. 6:199-222. Jorgenson, S.E., 1993. Modelling in ecotoxicology. In: A.K. Jakeman, M.B. Beck, and M.J. McAleer (Eds.), Modelling Change in Environmental Systems, Wiley, New York, 293-316. Jorgenson, S.E., 1995. Modeling toxic contaminants in an aquatic environment. In: V. Novotny and L. Somlyody (Eds.), Remediation and Management of Degraded River Basins. Springer, Berlin, 157-195. Kabata-Pendias, A. and H. Pendias, 1991. Trace elements in soils and plants. Boca Raton, CRC Press, 1-35. Kadlec, R.H., and Knight, R.L., 1996. Treatment Wetlands; CRC Press: New York, 241-274. Karlson, U., and W.T. Frankenberger, 1988. Dissipation of Soil Selenium By Microbial Volatilization At Kesterson Reservoir. Journal of the Soil Science Society of America. 52: 1640-1647. Karlson, U., and W.T. Frankenberger, 1990. Study of the Environmental Factors and a Possible Measures for Detoxification of Soil; evaluation of microbial volatilization of selenium. Science of the Total Environment. 92: 41-44. Klemm, D.J., P.A. Lewis, F. Fulk, and J.M. Lazorchak, 1990. Macroinvertebrate field and laboratory methods for evaluating the biological integrity of surface waters. Environmental Monitoring Systems Laboratory, U.S. Environmental Protection Agency, Cincinatti, OH 45268. EPA/600/4-90/030. Lakin, H.W. and D.F. Dawidson, 1967. The relation of the geochemistry of selenium to its occurrence in soil. In: Proceedings of Selenium in Biomedicine, Westport, Conn., 27-34. Lemly, A.D., 1987. Aquatic cycling of selenium: Implications for fish and wildlife. U.S. Department of the Interior. Fish and Wildlife Service Leaflet, No.12. Washington, D.C. Lortie, L, W.D. Gould, S. Rajan, R.G.L. McCready, and K.J. Cheng, 1992. Reduction of selenate and selenite to elemental selenium by Pseudomonas stutzeri isolate. Applied Environmental Microbiology. 58: 4042-4044. Macy, J.M., T.A. Michel, and D.G. Kirsch, 1989. Selenate reduction by a Pseudomonas species: a new model of anaerobic respiration. FEMS Microbiology Letters. 6 1 : 195-198. Macy, J.M., and S. Rech, 1990. The selenate reductase and nitrate reductase of a selenate- and nitrate-respiring Pseudomonas sp. are different enzymes, 90 t h Annual Meeting of the American Society of Microbiology, Anaheim, Calif., 241-257. Maiers, D.T., P.L. Wichlacz, D.L. Thompson, and D.F. Bruhn, 1988. Precipitation of Se°from selenate was reported in cultures recovered from the Kesterson National Wildlife Refuge. Applied Environmental Microbiology. 54: 2591-2593. Masscheleyn, P.H., and W.H. Patrick, 1993. Biogeochemical processes affecting selenium cycling in wetlands. Environmental Toxicology and Chemistry, 12: 2235-2243. 156 McDonald, L.E., and M.M. Strosher, 1998. Selenium mobilization from surface coal mining in the Elk River Basin, British Columbia. Cranbrook, British Columbia. BC Ministry of Environment, Lands, and Parks: 56. Millenium Excalibur User Manual, 1997. P S Analytical Ltd. Arthur House, Unit 3 Crayfields Industrial Estate, Main Road, St Pauls Cray, Orpington, Kent. BR5 3H0. UK. Nagpal, N.K., L.W. Pommen, and L.G. Swain, 1995. Approved and Working Criteria for Water Quality - 1995. Ministry of Environment, Lands, and Parks. Water Quality Branch, Victoria, B.C. 45 p. Neal, R.H., G. Sposito, K.M. Holtzclaw, and S.J. Traina, 1987. Selenite adsorption on alluvial soils, I, soil composition and pH effects. Journal of the Soil Science Society of America. 51 : 1161-1165. Neal, R.H., and G. Sposito, 1989. Selenate adsorption on alluvial soils. Journal of the Soil Science Society of America. 53, 70-74. Newland, L.W., 1982. Handbook of Environmental Chemistry, Springer-Verlag, New York, pp.45-47. Odum, H.T., 2000. Heavy metals in the environment: using wetlands for their removal. CRC Press, Boca Raton, Florida, 1-75. Oremland, R.S., J.T. Hollibaugh, A.S. Maest, T.S. Presser, L.G. Miller, and C.W. Culbertson, 1989. Selenate reduction to elemental selenium by anaerobic bacteria in sediments and culture. Applied Environmental Microbiology. 55:2333. Oremland, R.S. 1994. Biogeochemical Transformations of Selenium in Anoxic Environments. In Selenium in the Environment. Ed. W.T. Frankenberger. 1994. Marcel Dekker, New York. Oremland, R.S., J.S. Blum, C.W. Culbertson, P.T. Visscher, L.G. Miller, P. Dowdle, and F.E. Strohmaier, 1994. Isolation, Growth, and Metabolism of an Obligately Anaerobic, Selenate-Respiring Bacterium, Strain SES-3. Applied and Environmental Microbiology. Pp. 3011-3019. Paasikallio, A., 1981. The effect of soil pH and Fe on the availability of Se-75 in sphagnum peat soil. Annual Agriculture Fenn. 20:15-18. Rashid, K., and H.R. Krouse, 1985. The use of Se isotopes to demonstrate chemical selenite reduction. Canadian Journal of Chemistry. 63:3195-3199. Reamer, D.C., and W.H. Zoller,1980. Microbial transformation of selenium in aerobic sewage contaminated with elemental selenium and selenite. Science. 208: 500-508. Rech, S., and Macy, J.M., 1992. The terminal reductases for selenate and nitrate respiration in Thauera selenatis are two distinct enzymes. Journal of Bacteriology. 174: 7316-7320. Sposito, G., J.C.M De Wit, and R.H. Neal, 1988. Selenite adsorption on alluvial soils. Ill Chemical modeling, Journal of the Soil Science Society of America. 52; 947-950. Sposito, G., 1989. The Chemistry of Soils. Oxford University Press, New York, 148-169. Steinberg, N.A., J. Switzer Blum, L. Hochstein, and R.S. Oremland, 1992. Nitrate as a preferred electron acceptor for growth of selenate-respiring bacteria. Applied Environmental Microbiology. 58: 426-428. 157 Stork, A., W.A. Jury, and W.T. Frankemberger, 1999. Accelerated volatilization rates of selenium from different soils. Biological Trace Element Research. 69: 217-234. Stroiazzo, J., 1999. Molybdenum treatment and water management at Brenda Mines. Proceedings of the 1999 workshop: Molybdenum issues in Reclamation. The British Columbia Technical and Research Committee on Reclamation Thoman, R.V., 1984. Physico-chemical and ecological modeling the fate of toxic substances in natural water systems. Ecological Modelling, 22(1): 45-70. Thompson-Eagle, E.T., and W.T. Frankenberger, 1990. Site Volatilization of Selenium with Alternative Sources of Protein for Microbial Deselenification at Evaporation Ponds. Journal of Environmental Quality. 19: 125-129. Tomei, F.A., L.L. Barton, C L . Lemanski, and T.G. Zocco, 1992. Reduction of selenate and selenite to elemental selenium by Wolinella succinogenes. Canadian Journal of Microbiology. 38: 1328-1333. USEPA, 1987. Ambient water quality criteria for selenium. EPA Criteria and Standards Division. EPA/440/5-87/006. 130p. USEPA, 1998. Peer Consultation Workshop on Selenium Aquatic Toxicity and Bioaccumulation. EPA Office of Water. Internet:: http://www.epa.gov/ost/selenium/report.html. Van Muylder and Pourbaix, 1994. Selenium, Commision des Etudes Fondamentales et Applications of CEBELCOR: pp. 554-559. Weerasooriya, S.V.R., S.B. Bulumulla, S.A. Tilakaratne Bandara, and M.U. Jayasekara, 1989. The pH dependence of selenite adsorption by goethite. International Journal of Environmental Studies. 33: 111-116. Widdel, F and N. Pfenning, 1984. Dissimilatory sulphate- or sulphur- reducing bacteria, p 663 -679. In N.R. Krieg (ed), Bergey's Manual of Systematic Bacteriology. Vol. 1. Williams & Wilkins Co., Baltimore. Zhang, Y. and J.N. Moore. 1997a. Environmental conditions controlling selenium volatilization from a wetland system. Environmental Science and Technology. 31 : 511-517. Zhang, Y. and J.N. Moore. 1997b. Interaction of selenate with a wetland sediment. Applied Geochemistry. 12: 685-691. Zhang, Y. and W.T. Frankenberger, 1999. Effects of soil moisture, depth, and organic amendments on selenium volatilization. Journal of Environmental Quality. 28: 1321-1326. Zhang, Y. and D.L. Sparks, 1990. Kinetics of selenate and selenite adsorption/desorption at the goethite/water interface. Environmental Science and Technology. 24: 1848-1856. Zehr, J.P., and R.S. Oremland, 1987. Reduction of selenate toselenide by sulphate-respiring bacteria. Applied Environmental Microbiology. 53:1365-1369. Zieve, R., and P.J. Peterson, 1981. Volatilization of selenium from plants and soils. Science of the Total Environment. 19: 277-283. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0090217/manifest

Comment

Related Items